Edit on GitHub

Semantic Diff for SQL

by Iaroslav Zeigerman

Motivation

Software is constantly changing and evolving, and identifying what has changed and reviewing those changes is an integral part of the development process. SQL code is no exception to this.

Text-based diff tools such as git diff, when applied to a code base, have certain limitations. First, they can only detect insertions and deletions, not movements or updates of individual pieces of code. Second, such tools can only detect changes between lines of text, which is too coarse for something as granular and detailed as source code. Additionally, the outcome of such a diff is dependent on the underlying code formatting, and yields different results if the formatting should change.

Consider the following diff generated by Git:

Git diff output

Semantically the query hasn’t changed. The two arguments b and c have been swapped (moved), posing no impact on the output of the query. Yet Git replaced the whole affected expression alongside a bulk of unrelated elements.

The alternative to text-based diffing is to compare Abstract Syntax Trees (AST) instead. The main advantage of ASTs are that they are a direct product of code parsing, which represents the underlying code structure at any desired level of granularity. Comparing ASTs may yield extremely precise diffs; changes such as code movements and updates can also be detected. Even more importantly, this approach facilitates additional use cases beyond eyeballing two versions of source code side by side.

The use cases I had in mind for SQL when I decided to embark on this journey of semantic diffing were the following:

  • Query similarity score. Identifying which parts the two queries have in common to automatically suggest opportunities for consolidation, creation of intermediate/staging tables, and so on.
  • Differentiating between cosmetic / structural changes and functional ones. For example when a nested query is refactored into a common table expression (CTE), this kind of change doesn’t have any functional impact on either a query or its outcome.
  • Automatic suggestions about the need to retroactively backfill data. This is especially important for pipelines that populate very large tables for which restatement is a runtime-intensive procedure. The ability to discern between simple code movements and actual modifications can help assess the impact of a change and make suggestions accordingly.

The implementation discussed in this post is now a part of the SQLGlot library. You can find a complete source code in the diff.py module. The choice of SQLglot was an obvious one due to its simple but powerful API, lack of external dependencies and, more importantly, extensive list of supported SQL dialects.

The Search for a Solution

When it comes to any diffing tool (not just a semantic one), the primary challenge is to match as many elements of compared entities as possible. Once such a set of matching elements is available, deriving a sequence of changes becomes an easy task.

If our elements have unique identifiers associated with them (for example, an element’s ID in DOM), the matching problem is trivial. However, the SQL syntax trees that we are comparing have neither unique keys nor object identifiers that can be used for the purposes of matching. So, how do we suppose to find pairs of nodes that are related?

To better illustrate the problem, consider comparing the following SQL expressions: SELECT a + b + c, d, e and SELECT a - b + c, e, f. Matching individual nodes from respective syntax trees can be visualized as follows:

Figure 1: Example of node matching for two SQL expression trees Figure 1: Example of node matching for two SQL expression trees.

By looking at the figure of node matching for two SQL expression trees above, we conclude that the following changes should be captured by our solution:

  • Inserted nodes: Sub and f. These are the nodes from the target AST which do not have a matching node in the source AST.
  • Removed nodes: Add and d. These are the nodes from the source AST which do not have a counterpart in the target AST.
  • Remaining nodes must be identified as unchanged.

It should be clear at this point that if we manage to match nodes in the source tree with their counterparts in the target tree, then computing the diff becomes a trivial matter.

Naïve Brute-Force

The naïve solution would be to try all different permutations of node pair combinations, and see which set of pairs performs the best based on some type of heuristics. The runtime cost of such a solution quickly reaches the escape velocity; if both trees had only 10 nodes each, the number of such sets would approximately be 10! ^ 2 = 3.6M ^ 2 ~= 13 * 10^12. This is a very bad case of factorial complexity (to be precise, it’s actually much worse - O(n! ^ 2) - but I couldn’t come up with a name for it), so there is little need to explore this approach any further.

Myers Algorithm

After the naïve approach was proven to be infeasible, the next question I asked myself was “how does git diff work?”. This question led me to discover the Myers diff algorithm [1]. This algorithm has been designed to compare sequences of strings. At its core, it’s looking for the shortest path on a graph of possible edits that transform the first sequence into the second one, while heavily rewarding those paths that lead to longest subsequences of unchanged elements. There’s a lot of material out there describing this algorithm in greater detail. I found James Coglan’s series of blog posts to be the most comprehensive.

Therefore, I had this “brilliant” (actually not) idea to transform trees into sequences by traversing them in topological order, and then applying the Myers algorithm on resulting sequences while using a custom heuristics when checking the equality of two nodes. Unsurprisingly, comparing sequences of strings is quite different from comparing hierarchical tree structures, and by flattening trees into sequences, we lose a lot of relevant context. This resulted in a terrible performance of this algorithm on ASTs. It often matched completely unrelated nodes, even when the two trees were mostly the same, and produced extremely inaccurate lists of changes overall. After playing around with it a little and tweaking my equality heuristics to improve accuracy, I ultimately scrapped the whole implementation and went back to the drawing board.

Change Distiller

The algorithm I settled on at the end was Change Distiller, created by Fluri et al. [2], which in turn is an improvement over the core idea described by Chawathe et al. [3].

The algorithm consists of two high-level steps:

  1. Finding appropriate matchings between pairs of nodes that are part of compared ASTs. Identifying what is meant by “appropriate” matching is also a part of this step.
  2. Generating the so-called “edit script” from the matching set built in the 1st step. The edit script is a sequence of edit operations (for example, insert, remove, update, etc.) on individual tree nodes, such that when applied as transformations on the source AST, it eventually becomes the target AST. In general, the shorter the sequence, the better. The length of the edit script can be used to compare the performance of different algorithms, though this is not the only metric that matters.

The rest of this section is dedicated to the Python implementation of the steps above using the AST implementation provided by the SQLGlot library.

Building the Matching Set

Matching Leaves

We begin composing the matching set by matching the leaf nodes. Leaf nodes are the nodes that do not have any children nodes (such as literals, identifiers, etc.). In order to match them, we gather all the leaf nodes from the source tree and generate a cartesian product with all the leaves from the target tree, while comparing pairs created this way and assigning them a similarity score. During this stage, we also exclude pairs that don’t pass basic matching criteria. Then, we pick pairs that scored the highest while making sure that each node is matched no more than once.

Using the example provided at the beginning of the post, the process of building an initial set of candidate matchings can be seen on Figure 2.

Figure 2: Building a set of candidate matchings between leaf nodes. The third item in each triplet represents a similarity score between two nodes. Figure 2: Building a set of candidate matchings between leaf nodes. The third item in each triplet represents a similarity score between two nodes.

First, let’s analyze the similarity score. Then, we’ll discuss matching criteria.

The similarity score proposed by Fluri et al. [2] is a dice coefficient applied to bigrams of respective node values. A bigram is a sequence of two adjacent elements from a string computed in a sliding window fashion:

def bigram(string):
    count = max(0, len(string) - 1)
    return [string[i : i + 2] for i in range(count)]

For reasons that will become clear shortly, we actually need to compute bigram histograms rather than just sequences:

from collections import defaultdict

def bigram_histo(string):
    count = max(0, len(string) - 1)
    bigram_histo = defaultdict(int)
    for i in range(count):
        bigram_histo[string[i : i + 2]] += 1
    return bigram_histo

The dice coefficient formula looks like following:

Dice Coefficient

Where X is a bigram of the source node and Y is a bigram of the second one. What this essentially does is count the number of bigram elements the two nodes have in common, multiply it by 2, and then divide by the total number of elements in both bigrams. This is where bigram histograms come in handy:

def dice_coefficient(source, target):
    source_histo = bigram_histo(source.sql())
    target_histo = bigram_histo(target.sql())

    total_grams = (
        sum(source_histo.values()) + sum(target_histo.values())
    )
    if not total_grams:
        return 1.0 if source == target else 0.0

    overlap_len = 0
    overlapping_grams = set(source_histo) & set(target_histo)
    for g in overlapping_grams:
        overlap_len += min(source_histo[g], target_histo[g])

    return 2 * overlap_len / total_grams

To compute a bigram given a tree node, we first transform the node into its canonical SQL representation,so that the Literal(123) node becomes just “123” and the Identifier(“a”) node becomes just “a”. We also handle a scenario when strings are too short to derive bigrams. In this case, we fallback to checking the two nodes for equality.

Now when we know how to compute the similarity score, we can take care of the matching criteria for leaf nodes. In the original paper [2], the matching criteria is formalized as follows:

Matching criteria for leaf nodes

The two nodes are matched if two conditions are met:

  1. The node labels match (in our case labels are just node types).
  2. The similarity score for node values is greater than or equal to some threshold “f”. The authors of the paper recommend setting the value of “f” to 0.6.

With building blocks in place, we can now build a matching set for leaf nodes. First, we generate a list of candidates for matching:

from heapq import heappush, heappop

candidate_matchings = []
source_leaves = _get_leaves(self._source)
target_leaves = _get_leaves(self._target)
for source_leaf in source_leaves:
    for target_leaf in target_leaves:
        if _is_same_type(source_leaf, target_leaf):
            similarity_score = dice_coefficient(
                source_leaf, target_leaf
            )
            if similarity_score >= 0.6:
                heappush(
                    candidate_matchings,
                    (
                        -similarity_score,
                        len(candidate_matchings),
                        source_leaf,
                        target_leaf,
                    ),
                )

In the implementation above, we push each matching pair onto the heap to automatically maintain the correct order based on the assigned similarity score.

Finally, we build the initial matching set by picking leaf pairs with the highest score:

matching_set = set()
while candidate_matchings:
    _, _, source_leaf, target_leaf = heappop(candidate_matchings)
    if (
        source_leaf in unmatched_source_nodes
        and target_leaf in unmatched_target_nodes
    ):
        matching_set.add((source_leaf, target_leaf))
        unmatched_source_nodes.remove(source_leaf)
        unmatched_target_nodes.remove(target_leaf)

To finalize the matching set, we should now proceed with matching inner nodes.

Matching Inner Nodes

Matching inner nodes is quite similar to matching leaf nodes, with the following two distinctions:

  • Rather than ranking a set of possible candidates, we pick the first node pair that passes the matching criteria.
  • The matching criteria itself has been extended to account for the number of leaf nodes the pair of inner nodes have in common.

Figure 3: Matching inner nodes based on their type as well as how many of their leaf nodes have been previously matched. Figure 3: Matching inner nodes based on their type as well as how many of their leaf nodes have been previously matched.

Let’s start with the matching criteria. The criteria is formalized as follows:

Matching criteria for inner nodes

Alongside already familiar similarity score and node type criteria, there is a new one in the middle: the ratio of leaf nodes that the two nodes have in common must exceed some threshold “t”. The recommended value for “t” is also 0.6. Counting the number of common leaf nodes is pretty straightforward, since we already have the complete matching set for leaves. All we need to do is count how many matching pairs do leaf nodes from the two compared inner nodes form.

There are two additional heuristics associated with this matching criteria:

  • Inner node similarity weighting: if the similarity score between the node values doesn’t pass the threshold “f” but the ratio of common leaf nodes (“t”) is greater than or equal to 0.8, then the matching is considered successful.
  • The threshold “t” is reduced to 0.4 for inner nodes with the number of leaf nodes equal to 4 or less, in order to decrease the false negative rate for small subtrees.

We now only have to iterate through the remaining unmatched nodes and form matching pairs based on the outlined criteria:

leaves_matching_set = matching_set.copy()

for source_node in unmatched_source_nodes.copy():
    for target_node in unmatched_target_nodes:
        if _is_same_type(source_node, target_node):
            source_leaves = set(_get_leaves(source_node))
            target_leaves = set(_get_leaves(target_node))

            max_leaves_num = max(len(source_leaves), len(target_leaves))
            if max_leaves_num:
                common_leaves_num = sum(
                    1 if s in source_leaves and t in target_leaves else 0
                    for s, t in leaves_matching_set
                )
                leaf_similarity_score = common_leaves_num / max_leaves_num
            else:
                leaf_similarity_score = 0.0

            adjusted_t = (
                0.6
                if min(len(source_leaves), len(target_leaves)) > 4
                else 0.4
            )

            if leaf_similarity_score >= 0.8 or (
                leaf_similarity_score >= adjusted_t
                and dice_coefficient(source_node, target_node) >= 0.6
            ):
                matching_set.add((source_node, target_node))
                unmatched_source_nodes.remove(source_node)
                unmatched_target_nodes.remove(target_node)
                break

After the matching set is formed, we can proceed with generation of the edit script, which will be the algorithm’s output.

Generating the Edit Script

At this point, we should have the following 3 sets at our disposal:

  • The set of matched node pairs.
  • The set of remaining unmatched nodes from the source tree.
  • The set of remaining unmatched nodes from the target tree.

We can derive 3 kinds of edits from the matching set: either the node’s value was updated (Update), the node was moved to a different position within the tree (Move), or the node remained unchanged (Keep). Note that the Move case is not mutually exclusive with the other two. The node could have been updated or could have remained the same while at the same time its position within its parent node or the parent node itself could have changed. All unmatched nodes from the source tree are the ones that were removed (Remove), while unmatched nodes from the target tree are the ones that were inserted (Insert).

The latter two cases are pretty straightforward to implement:

edit_script = []

for removed_node in unmatched_source_nodes:
    edit_script.append(Remove(removed_node))
for inserted_node in unmatched_target_nodes:
    edit_script.append(Insert(inserted_node))

Traversing the matching set requires a little more thought:

for source_node, target_node in matching_set:
    if (
        not isinstance(source_node, LEAF_EXPRESSION_TYPES)
        or source_node == target_node
    ):
        move_edits = generate_move_edits(
            source_node, target_node, matching_set
        )
        edit_script.extend(move_edits)
        edit_script.append(Keep(source_node, target_node))
    else:
        edit_script.append(Update(source_node, target_node))

If a matching pair represents a pair of leaf nodes, we check if they are the same to decide whether an update took place. For inner node pairs, we also need to compare the positions of their respective children to detect node movements. Chawathe et al. [3] suggest applying the longest common subsequence (LCS) algorithm which, no surprise here, was described by Myers himself [1]. There is a small catch, however: instead of checking the equality of two children nodes, we need to check whether the two nodes form a pair that is a part of our matching set.

Now with this knowledge, the implementation becomes straightforward:

def generate_move_edits(source, target, matching_set):
    source_children = _get_child_nodes(source)
    target_children = _get_child_nodes(target)

    lcs = set(
        _longest_common_subsequence(
            source_children,
            target_children,
            lambda l, r: (l, r) in matching_set
        )
    )

    move_edits = []
    for node in source_children:
        if node not in lcs and node not in unmatched_source_nodes:
            move_edits.append(Move(node))

    return move_edits

I left out the implementation of the LCS algorithm itself here, but there are plenty of implementation choices out there that can be easily looked up.

Output

The implemented algorithm produces the output that resembles the following:

>>> from sqlglot import parse_one, diff
>>> diff(parse_one("SELECT a + b + c, d, e"), parse_one("SELECT a - b + c, e, f"))

Remove(Add)
Remove(Column(d))
Remove(Identifier(d))
Insert(Sub)
Insert(Column(f))
Insert(Identifier(f))
Keep(Select, Select)
Keep(Add, Add)
Keep(Column(a), Column(a))
Keep(Identifier(a), Identifier(a))
Keep(Column(b), Column(b))
Keep(Identifier(b), Identifier(b))
Keep(Column(c), Column(c))
Keep(Identifier(c), Identifier(c))
Keep(Column(e), Column(e))
Keep(Identifier(e), Identifier(e))

Note that the output above is abbreviated. The string representation of actual AST nodes is significantly more verbose.

The implementation works especially well when coupled with the SQLGlot’s query optimizer which can be used to produce canonical representations of compared queries:

>>> schema={"t": {"a": "INT", "b": "INT", "c": "INT", "d": "INT"}}
>>> source = """
... SELECT 1 + 1 + a
... FROM t
... WHERE b = 1 OR (c = 2 AND d = 3)
... """
>>> target = """
... SELECT 2 + a
... FROM t
... WHERE (b = 1 OR c = 2) AND (b = 1 OR d = 3)
... """
>>> optimized_source = optimize(parse_one(source), schema=schema)
>>> optimized_target = optimize(parse_one(target), schema=schema)
>>> edit_script = diff(optimized_source, optimized_target)
>>> sum(0 if isinstance(e, Keep) else 1 for e in edit_script)
0

Optimizations

The worst case runtime complexity of this algorithm is not exactly stellar: O(n^2 * log n^2). This is because of the leaf matching process, which involves ranking a cartesian product between all leaf nodes of compared trees. Unsurprisingly, the algorithm takes a considerable time to finish for bigger queries.

There are still a few basic things we can do in our implementation to help improve performance:

  • Refer to individual node objects using their identifiers (Python’s id()) instead of direct references in sets. This helps avoid costly recursive hash calculations and equality checks.
  • Cache bigram histograms to avoid computing them more than once for the same node.
  • Compute the canonical SQL string representation for each tree once while caching string representations of all inner nodes. This prevents redundant tree traversals when bigrams are computed.

At the time of writing only the first two optimizations have been implemented, so there is an opportunity to contribute for anyone who’s interested.

Alternative Solutions

This section is dedicated to solutions that I’ve investigated, but haven’t tried.

First, this section wouldn’t be complete without Tristan Hume’s blog post. Tristan’s solution has a lot in common with the Myers algorithm plus heuristics that is much more clever than what I came up with. The implementation relies on a combination of dynamic programming and A* search algorithm to explore the space of possible matchings and pick the best ones. It seemed to have worked well for Tistan’s specific use case, but after my negative experience with the Myers algorithm, I decided to try something different.

Another notable approach is the Gumtree algorithm by Falleri et al. [4]. I discovered this paper after I’d already implemented the algorithm that is the main focus of this post. In sections 5.2 and 5.3 of their paper, the authors compare the two algorithms side by side and claim that Gumtree is significantly better in terms of both runtime performance and accuracy when evaluated on 12 792 pairs of Java source files. This doesn’t surprise me, as the algorithm takes the height of subtrees into account. In my tests, I definitely saw scenarios in which this context would have helped. On top of that, the authors promise O(n^2) runtime complexity in the worst case which, given the Change Distiller's O(n^2 * log n^2), looks particularly tempting. I hope to try this algorithm out at some point, and there is a good chance you see me writing about it in my future posts.

Conclusion

The Change Distiller algorithm yielded quite satisfactory results in most of my tests. The scenarios in which it fell short mostly concerned identical (or very similar) subtrees located in different parts of the AST. In those cases, node mismatches were frequent and, as a result, edit scripts were somewhat suboptimal.

Additionally, the runtime performance of the algorithm leaves a lot to be desired. On trees with 1000 leaf nodes each, the algorithm takes a little under 2 seconds to complete. My implementation still has room for improvement, but this should give you a rough idea of what to expect. It appears that the Gumtree algorithm [4] can help address both of these points. I hope to find bandwidth to work on it soon and then compare the two algorithms side-by-side to find out which one performs better on SQL specifically. In the meantime, Change Distiller definitely gets the job done, and I can now proceed with applying it to some of the use cases I mentioned at the beginning of this post.

I’m also curious to learn whether other folks in the industry faced a similar problem, and how they approached it. If you did something similar, I’m interested to hear about your experience.

References

[1] Eugene W. Myers. An O(ND) Difference Algorithm and Its Variations. Algorithmica 1(2): 251-266 (1986)

[2] B. Fluri, M. Wursch, M. Pinzger, and H. Gall. Change Distilling: Tree differencing for fine-grained source code change extraction. IEEE Trans. Software Eng., 33(11):725–743, 2007.

[3] S.S. Chawathe, A. Rajaraman, H. Garcia-Molina, and J. Widom. Change Detection in Hierarchically Structured Information. Proc. ACM Sigmod Int’l Conf. Management of Data, pp. 493-504, June 1996

[4] Jean-Rémy Falleri, Floréal Morandat, Xavier Blanc, Matias Martinez, Martin Monperrus. Fine-grained and Accurate Source Code Differencing. Proceedings of the International Conference on Automated Software Engineering, 2014, Västeras, Sweden. pp.313-324, 10.1145/2642937.2642982. hal-01054552


  1"""
  2.. include:: ../posts/sql_diff.md
  3
  4----
  5"""
  6
  7from __future__ import annotations
  8
  9import typing as t
 10from collections import defaultdict
 11from dataclasses import dataclass
 12from heapq import heappop, heappush
 13
 14from sqlglot import Dialect, expressions as exp
 15from sqlglot.helper import seq_get
 16
 17if t.TYPE_CHECKING:
 18    from sqlglot.dialects.dialect import DialectType
 19
 20
 21@dataclass(frozen=True)
 22class Insert:
 23    """Indicates that a new node has been inserted"""
 24
 25    expression: exp.Expression
 26
 27
 28@dataclass(frozen=True)
 29class Remove:
 30    """Indicates that an existing node has been removed"""
 31
 32    expression: exp.Expression
 33
 34
 35@dataclass(frozen=True)
 36class Move:
 37    """Indicates that an existing node's position within the tree has changed"""
 38
 39    expression: exp.Expression
 40
 41
 42@dataclass(frozen=True)
 43class Update:
 44    """Indicates that an existing node has been updated"""
 45
 46    source: exp.Expression
 47    target: exp.Expression
 48
 49
 50@dataclass(frozen=True)
 51class Keep:
 52    """Indicates that an existing node hasn't been changed"""
 53
 54    source: exp.Expression
 55    target: exp.Expression
 56
 57
 58if t.TYPE_CHECKING:
 59    from sqlglot._typing import T
 60
 61    Edit = t.Union[Insert, Remove, Move, Update, Keep]
 62
 63
 64def diff(
 65    source: exp.Expression,
 66    target: exp.Expression,
 67    matchings: t.List[t.Tuple[exp.Expression, exp.Expression]] | None = None,
 68    delta_only: bool = False,
 69    copy: bool = True,
 70    **kwargs: t.Any,
 71) -> t.List[Edit]:
 72    """
 73    Returns the list of changes between the source and the target expressions.
 74
 75    Examples:
 76        >>> diff(parse_one("a + b"), parse_one("a + c"))
 77        [
 78            Remove(expression=(COLUMN this: (IDENTIFIER this: b, quoted: False))),
 79            Insert(expression=(COLUMN this: (IDENTIFIER this: c, quoted: False))),
 80            Keep(
 81                source=(ADD this: ...),
 82                target=(ADD this: ...)
 83            ),
 84            Keep(
 85                source=(COLUMN this: (IDENTIFIER this: a, quoted: False)),
 86                target=(COLUMN this: (IDENTIFIER this: a, quoted: False))
 87            ),
 88        ]
 89
 90    Args:
 91        source: the source expression.
 92        target: the target expression against which the diff should be calculated.
 93        matchings: the list of pre-matched node pairs which is used to help the algorithm's
 94            heuristics produce better results for subtrees that are known by a caller to be matching.
 95            Note: expression references in this list must refer to the same node objects that are
 96            referenced in source / target trees.
 97        delta_only: excludes all `Keep` nodes from the diff.
 98        copy: whether to copy the input expressions.
 99            Note: if this is set to false, the caller must ensure that there are no shared references
100            in the two ASTs, otherwise the diffing algorithm may produce unexpected behavior.
101        kwargs: additional arguments to pass to the ChangeDistiller instance.
102
103    Returns:
104        the list of Insert, Remove, Move, Update and Keep objects for each node in the source and the
105        target expression trees. This list represents a sequence of steps needed to transform the source
106        expression tree into the target one.
107    """
108    matchings = matchings or []
109    matching_ids = {id(n) for pair in matchings for n in pair}
110
111    def compute_node_mappings(
112        original: exp.Expression, copy: exp.Expression
113    ) -> t.Dict[int, exp.Expression]:
114        return {
115            id(old_node): new_node
116            for old_node, new_node in zip(original.walk(), copy.walk())
117            if id(old_node) in matching_ids
118        }
119
120    source_copy = source.copy() if copy else source
121    target_copy = target.copy() if copy else target
122
123    node_mappings = {
124        **compute_node_mappings(source, source_copy),
125        **compute_node_mappings(target, target_copy),
126    }
127    matchings_copy = [(node_mappings[id(s)], node_mappings[id(t)]) for s, t in matchings]
128
129    return ChangeDistiller(**kwargs).diff(
130        source_copy,
131        target_copy,
132        matchings=matchings_copy,
133        delta_only=delta_only,
134    )
135
136
137# The expression types for which Update edits are allowed.
138UPDATABLE_EXPRESSION_TYPES = (
139    exp.Alias,
140    exp.Boolean,
141    exp.Column,
142    exp.DataType,
143    exp.Lambda,
144    exp.Literal,
145    exp.Table,
146    exp.Window,
147)
148
149IGNORED_LEAF_EXPRESSION_TYPES = (exp.Identifier,)
150
151
152class ChangeDistiller:
153    """
154    The implementation of the Change Distiller algorithm described by Beat Fluri and Martin Pinzger in
155    their paper https://ieeexplore.ieee.org/document/4339230, which in turn is based on the algorithm by
156    Chawathe et al. described in http://ilpubs.stanford.edu:8090/115/1/1995-46.pdf.
157    """
158
159    def __init__(self, f: float = 0.6, t: float = 0.6, dialect: DialectType = None) -> None:
160        self.f = f
161        self.t = t
162        self._sql_generator = Dialect.get_or_raise(dialect).generator()
163
164    def diff(
165        self,
166        source: exp.Expression,
167        target: exp.Expression,
168        matchings: t.List[t.Tuple[exp.Expression, exp.Expression]] | None = None,
169        delta_only: bool = False,
170    ) -> t.List[Edit]:
171        matchings = matchings or []
172        pre_matched_nodes = {id(s): id(t) for s, t in matchings}
173        if len({n for pair in pre_matched_nodes.items() for n in pair}) != 2 * len(matchings):
174            raise ValueError("Each node can be referenced at most once in the list of matchings")
175
176        self._source = source
177        self._target = target
178        self._source_index = {
179            id(n): n for n in self._source.bfs() if not isinstance(n, IGNORED_LEAF_EXPRESSION_TYPES)
180        }
181        self._target_index = {
182            id(n): n for n in self._target.bfs() if not isinstance(n, IGNORED_LEAF_EXPRESSION_TYPES)
183        }
184        self._unmatched_source_nodes = set(self._source_index) - set(pre_matched_nodes)
185        self._unmatched_target_nodes = set(self._target_index) - set(pre_matched_nodes.values())
186        self._bigram_histo_cache: t.Dict[int, t.DefaultDict[str, int]] = {}
187
188        matching_set = self._compute_matching_set() | set(pre_matched_nodes.items())
189        return self._generate_edit_script(matching_set, delta_only)
190
191    def _generate_edit_script(
192        self,
193        matching_set: t.Set[t.Tuple[int, int]],
194        delta_only: bool,
195    ) -> t.List[Edit]:
196        edit_script: t.List[Edit] = []
197        for removed_node_id in self._unmatched_source_nodes:
198            edit_script.append(Remove(self._source_index[removed_node_id]))
199        for inserted_node_id in self._unmatched_target_nodes:
200            edit_script.append(Insert(self._target_index[inserted_node_id]))
201        for kept_source_node_id, kept_target_node_id in matching_set:
202            source_node = self._source_index[kept_source_node_id]
203            target_node = self._target_index[kept_target_node_id]
204
205            if (
206                not isinstance(source_node, UPDATABLE_EXPRESSION_TYPES)
207                or source_node == target_node
208            ):
209                edit_script.extend(
210                    self._generate_move_edits(source_node, target_node, matching_set)
211                )
212
213                source_non_expression_leaves = dict(_get_non_expression_leaves(source_node))
214                target_non_expression_leaves = dict(_get_non_expression_leaves(target_node))
215
216                if source_non_expression_leaves != target_non_expression_leaves:
217                    edit_script.append(Update(source_node, target_node))
218                elif not delta_only:
219                    edit_script.append(Keep(source_node, target_node))
220            else:
221                edit_script.append(Update(source_node, target_node))
222
223        return edit_script
224
225    def _generate_move_edits(
226        self, source: exp.Expression, target: exp.Expression, matching_set: t.Set[t.Tuple[int, int]]
227    ) -> t.List[Move]:
228        source_args = [id(e) for e in _expression_only_args(source)]
229        target_args = [id(e) for e in _expression_only_args(target)]
230
231        args_lcs = set(_lcs(source_args, target_args, lambda l, r: (l, r) in matching_set))
232
233        move_edits = []
234        for a in source_args:
235            if a not in args_lcs and a not in self._unmatched_source_nodes:
236                move_edits.append(Move(self._source_index[a]))
237
238        return move_edits
239
240    def _compute_matching_set(self) -> t.Set[t.Tuple[int, int]]:
241        leaves_matching_set = self._compute_leaf_matching_set()
242        matching_set = leaves_matching_set.copy()
243
244        ordered_unmatched_source_nodes = {
245            id(n): None for n in self._source.bfs() if id(n) in self._unmatched_source_nodes
246        }
247        ordered_unmatched_target_nodes = {
248            id(n): None for n in self._target.bfs() if id(n) in self._unmatched_target_nodes
249        }
250
251        for source_node_id in ordered_unmatched_source_nodes:
252            for target_node_id in ordered_unmatched_target_nodes:
253                source_node = self._source_index[source_node_id]
254                target_node = self._target_index[target_node_id]
255                if _is_same_type(source_node, target_node):
256                    source_leaf_ids = {id(l) for l in _get_expression_leaves(source_node)}
257                    target_leaf_ids = {id(l) for l in _get_expression_leaves(target_node)}
258
259                    max_leaves_num = max(len(source_leaf_ids), len(target_leaf_ids))
260                    if max_leaves_num:
261                        common_leaves_num = sum(
262                            1 if s in source_leaf_ids and t in target_leaf_ids else 0
263                            for s, t in leaves_matching_set
264                        )
265                        leaf_similarity_score = common_leaves_num / max_leaves_num
266                    else:
267                        leaf_similarity_score = 0.0
268
269                    adjusted_t = (
270                        self.t if min(len(source_leaf_ids), len(target_leaf_ids)) > 4 else 0.4
271                    )
272
273                    if leaf_similarity_score >= 0.8 or (
274                        leaf_similarity_score >= adjusted_t
275                        and self._dice_coefficient(source_node, target_node) >= self.f
276                    ):
277                        matching_set.add((source_node_id, target_node_id))
278                        self._unmatched_source_nodes.remove(source_node_id)
279                        self._unmatched_target_nodes.remove(target_node_id)
280                        ordered_unmatched_target_nodes.pop(target_node_id, None)
281                        break
282
283        return matching_set
284
285    def _compute_leaf_matching_set(self) -> t.Set[t.Tuple[int, int]]:
286        candidate_matchings: t.List[t.Tuple[float, int, int, exp.Expression, exp.Expression]] = []
287        source_expression_leaves = list(_get_expression_leaves(self._source))
288        target_expression_leaves = list(_get_expression_leaves(self._target))
289        for source_leaf in source_expression_leaves:
290            for target_leaf in target_expression_leaves:
291                if _is_same_type(source_leaf, target_leaf):
292                    similarity_score = self._dice_coefficient(source_leaf, target_leaf)
293                    if similarity_score >= self.f:
294                        heappush(
295                            candidate_matchings,
296                            (
297                                -similarity_score,
298                                -_parent_similarity_score(source_leaf, target_leaf),
299                                len(candidate_matchings),
300                                source_leaf,
301                                target_leaf,
302                            ),
303                        )
304
305        # Pick best matchings based on the highest score
306        matching_set = set()
307        while candidate_matchings:
308            _, _, _, source_leaf, target_leaf = heappop(candidate_matchings)
309            if (
310                id(source_leaf) in self._unmatched_source_nodes
311                and id(target_leaf) in self._unmatched_target_nodes
312            ):
313                matching_set.add((id(source_leaf), id(target_leaf)))
314                self._unmatched_source_nodes.remove(id(source_leaf))
315                self._unmatched_target_nodes.remove(id(target_leaf))
316
317        return matching_set
318
319    def _dice_coefficient(self, source: exp.Expression, target: exp.Expression) -> float:
320        source_histo = self._bigram_histo(source)
321        target_histo = self._bigram_histo(target)
322
323        total_grams = sum(source_histo.values()) + sum(target_histo.values())
324        if not total_grams:
325            return 1.0 if source == target else 0.0
326
327        overlap_len = 0
328        overlapping_grams = set(source_histo) & set(target_histo)
329        for g in overlapping_grams:
330            overlap_len += min(source_histo[g], target_histo[g])
331
332        return 2 * overlap_len / total_grams
333
334    def _bigram_histo(self, expression: exp.Expression) -> t.DefaultDict[str, int]:
335        if id(expression) in self._bigram_histo_cache:
336            return self._bigram_histo_cache[id(expression)]
337
338        expression_str = self._sql_generator.generate(expression)
339        count = max(0, len(expression_str) - 1)
340        bigram_histo: t.DefaultDict[str, int] = defaultdict(int)
341        for i in range(count):
342            bigram_histo[expression_str[i : i + 2]] += 1
343
344        self._bigram_histo_cache[id(expression)] = bigram_histo
345        return bigram_histo
346
347
348def _get_expression_leaves(expression: exp.Expression) -> t.Iterator[exp.Expression]:
349    has_child_exprs = False
350
351    for node in expression.iter_expressions():
352        if not isinstance(node, IGNORED_LEAF_EXPRESSION_TYPES):
353            has_child_exprs = True
354            yield from _get_expression_leaves(node)
355
356    if not has_child_exprs:
357        yield expression
358
359
360def _get_non_expression_leaves(expression: exp.Expression) -> t.Iterator[t.Tuple[str, t.Any]]:
361    for arg, value in expression.args.items():
362        if isinstance(value, exp.Expression) or (
363            isinstance(value, list) and isinstance(seq_get(value, 0), exp.Expression)
364        ):
365            continue
366
367        yield (arg, value)
368
369
370def _is_same_type(source: exp.Expression, target: exp.Expression) -> bool:
371    if type(source) is type(target):
372        if isinstance(source, exp.Join):
373            return source.args.get("side") == target.args.get("side")
374
375        if isinstance(source, exp.Anonymous):
376            return source.this == target.this
377
378        return True
379
380    return False
381
382
383def _parent_similarity_score(
384    source: t.Optional[exp.Expression], target: t.Optional[exp.Expression]
385) -> int:
386    if source is None or target is None or type(source) is not type(target):
387        return 0
388
389    return 1 + _parent_similarity_score(source.parent, target.parent)
390
391
392def _expression_only_args(expression: exp.Expression) -> t.Iterator[exp.Expression]:
393    yield from (
394        arg
395        for arg in expression.iter_expressions()
396        if not isinstance(arg, IGNORED_LEAF_EXPRESSION_TYPES)
397    )
398
399
400def _lcs(
401    seq_a: t.Sequence[T], seq_b: t.Sequence[T], equal: t.Callable[[T, T], bool]
402) -> t.Sequence[t.Optional[T]]:
403    """Calculates the longest common subsequence"""
404
405    len_a = len(seq_a)
406    len_b = len(seq_b)
407    lcs_result = [[None] * (len_b + 1) for i in range(len_a + 1)]
408
409    for i in range(len_a + 1):
410        for j in range(len_b + 1):
411            if i == 0 or j == 0:
412                lcs_result[i][j] = []  # type: ignore
413            elif equal(seq_a[i - 1], seq_b[j - 1]):
414                lcs_result[i][j] = lcs_result[i - 1][j - 1] + [seq_a[i - 1]]  # type: ignore
415            else:
416                lcs_result[i][j] = (
417                    lcs_result[i - 1][j]
418                    if len(lcs_result[i - 1][j]) > len(lcs_result[i][j - 1])  # type: ignore
419                    else lcs_result[i][j - 1]
420                )
421
422    return lcs_result[len_a][len_b]  # type: ignore
@dataclass(frozen=True)
class Insert:
22@dataclass(frozen=True)
23class Insert:
24    """Indicates that a new node has been inserted"""
25
26    expression: exp.Expression

Indicates that a new node has been inserted

Insert(expression: sqlglot.expressions.Expression)
@dataclass(frozen=True)
class Remove:
29@dataclass(frozen=True)
30class Remove:
31    """Indicates that an existing node has been removed"""
32
33    expression: exp.Expression

Indicates that an existing node has been removed

Remove(expression: sqlglot.expressions.Expression)
@dataclass(frozen=True)
class Move:
36@dataclass(frozen=True)
37class Move:
38    """Indicates that an existing node's position within the tree has changed"""
39
40    expression: exp.Expression

Indicates that an existing node's position within the tree has changed

Move(expression: sqlglot.expressions.Expression)
@dataclass(frozen=True)
class Update:
43@dataclass(frozen=True)
44class Update:
45    """Indicates that an existing node has been updated"""
46
47    source: exp.Expression
48    target: exp.Expression

Indicates that an existing node has been updated

@dataclass(frozen=True)
class Keep:
51@dataclass(frozen=True)
52class Keep:
53    """Indicates that an existing node hasn't been changed"""
54
55    source: exp.Expression
56    target: exp.Expression

Indicates that an existing node hasn't been changed

def diff( source: sqlglot.expressions.Expression, target: sqlglot.expressions.Expression, matchings: Optional[List[Tuple[sqlglot.expressions.Expression, sqlglot.expressions.Expression]]] = None, delta_only: bool = False, copy: bool = True, **kwargs: Any) -> List[Union[Insert, Remove, Move, Update, Keep]]:
 65def diff(
 66    source: exp.Expression,
 67    target: exp.Expression,
 68    matchings: t.List[t.Tuple[exp.Expression, exp.Expression]] | None = None,
 69    delta_only: bool = False,
 70    copy: bool = True,
 71    **kwargs: t.Any,
 72) -> t.List[Edit]:
 73    """
 74    Returns the list of changes between the source and the target expressions.
 75
 76    Examples:
 77        >>> diff(parse_one("a + b"), parse_one("a + c"))
 78        [
 79            Remove(expression=(COLUMN this: (IDENTIFIER this: b, quoted: False))),
 80            Insert(expression=(COLUMN this: (IDENTIFIER this: c, quoted: False))),
 81            Keep(
 82                source=(ADD this: ...),
 83                target=(ADD this: ...)
 84            ),
 85            Keep(
 86                source=(COLUMN this: (IDENTIFIER this: a, quoted: False)),
 87                target=(COLUMN this: (IDENTIFIER this: a, quoted: False))
 88            ),
 89        ]
 90
 91    Args:
 92        source: the source expression.
 93        target: the target expression against which the diff should be calculated.
 94        matchings: the list of pre-matched node pairs which is used to help the algorithm's
 95            heuristics produce better results for subtrees that are known by a caller to be matching.
 96            Note: expression references in this list must refer to the same node objects that are
 97            referenced in source / target trees.
 98        delta_only: excludes all `Keep` nodes from the diff.
 99        copy: whether to copy the input expressions.
100            Note: if this is set to false, the caller must ensure that there are no shared references
101            in the two ASTs, otherwise the diffing algorithm may produce unexpected behavior.
102        kwargs: additional arguments to pass to the ChangeDistiller instance.
103
104    Returns:
105        the list of Insert, Remove, Move, Update and Keep objects for each node in the source and the
106        target expression trees. This list represents a sequence of steps needed to transform the source
107        expression tree into the target one.
108    """
109    matchings = matchings or []
110    matching_ids = {id(n) for pair in matchings for n in pair}
111
112    def compute_node_mappings(
113        original: exp.Expression, copy: exp.Expression
114    ) -> t.Dict[int, exp.Expression]:
115        return {
116            id(old_node): new_node
117            for old_node, new_node in zip(original.walk(), copy.walk())
118            if id(old_node) in matching_ids
119        }
120
121    source_copy = source.copy() if copy else source
122    target_copy = target.copy() if copy else target
123
124    node_mappings = {
125        **compute_node_mappings(source, source_copy),
126        **compute_node_mappings(target, target_copy),
127    }
128    matchings_copy = [(node_mappings[id(s)], node_mappings[id(t)]) for s, t in matchings]
129
130    return ChangeDistiller(**kwargs).diff(
131        source_copy,
132        target_copy,
133        matchings=matchings_copy,
134        delta_only=delta_only,
135    )

Returns the list of changes between the source and the target expressions.

Examples:
>>> diff(parse_one("a + b"), parse_one("a + c"))
[
    Remove(expression=(COLUMN this: (IDENTIFIER this: b, quoted: False))),
    Insert(expression=(COLUMN this: (IDENTIFIER this: c, quoted: False))),
    Keep(
        source=(ADD this: ...),
        target=(ADD this: ...)
    ),
    Keep(
        source=(COLUMN this: (IDENTIFIER this: a, quoted: False)),
        target=(COLUMN this: (IDENTIFIER this: a, quoted: False))
    ),
]
Arguments:
  • source: the source expression.
  • target: the target expression against which the diff should be calculated.
  • matchings: the list of pre-matched node pairs which is used to help the algorithm's heuristics produce better results for subtrees that are known by a caller to be matching. Note: expression references in this list must refer to the same node objects that are referenced in source / target trees.
  • delta_only: excludes all Keep nodes from the diff.
  • copy: whether to copy the input expressions. Note: if this is set to false, the caller must ensure that there are no shared references in the two ASTs, otherwise the diffing algorithm may produce unexpected behavior.
  • kwargs: additional arguments to pass to the ChangeDistiller instance.
Returns:

the list of Insert, Remove, Move, Update and Keep objects for each node in the source and the target expression trees. This list represents a sequence of steps needed to transform the source expression tree into the target one.

IGNORED_LEAF_EXPRESSION_TYPES = (<class 'sqlglot.expressions.Identifier'>,)
class ChangeDistiller:
153class ChangeDistiller:
154    """
155    The implementation of the Change Distiller algorithm described by Beat Fluri and Martin Pinzger in
156    their paper https://ieeexplore.ieee.org/document/4339230, which in turn is based on the algorithm by
157    Chawathe et al. described in http://ilpubs.stanford.edu:8090/115/1/1995-46.pdf.
158    """
159
160    def __init__(self, f: float = 0.6, t: float = 0.6, dialect: DialectType = None) -> None:
161        self.f = f
162        self.t = t
163        self._sql_generator = Dialect.get_or_raise(dialect).generator()
164
165    def diff(
166        self,
167        source: exp.Expression,
168        target: exp.Expression,
169        matchings: t.List[t.Tuple[exp.Expression, exp.Expression]] | None = None,
170        delta_only: bool = False,
171    ) -> t.List[Edit]:
172        matchings = matchings or []
173        pre_matched_nodes = {id(s): id(t) for s, t in matchings}
174        if len({n for pair in pre_matched_nodes.items() for n in pair}) != 2 * len(matchings):
175            raise ValueError("Each node can be referenced at most once in the list of matchings")
176
177        self._source = source
178        self._target = target
179        self._source_index = {
180            id(n): n for n in self._source.bfs() if not isinstance(n, IGNORED_LEAF_EXPRESSION_TYPES)
181        }
182        self._target_index = {
183            id(n): n for n in self._target.bfs() if not isinstance(n, IGNORED_LEAF_EXPRESSION_TYPES)
184        }
185        self._unmatched_source_nodes = set(self._source_index) - set(pre_matched_nodes)
186        self._unmatched_target_nodes = set(self._target_index) - set(pre_matched_nodes.values())
187        self._bigram_histo_cache: t.Dict[int, t.DefaultDict[str, int]] = {}
188
189        matching_set = self._compute_matching_set() | set(pre_matched_nodes.items())
190        return self._generate_edit_script(matching_set, delta_only)
191
192    def _generate_edit_script(
193        self,
194        matching_set: t.Set[t.Tuple[int, int]],
195        delta_only: bool,
196    ) -> t.List[Edit]:
197        edit_script: t.List[Edit] = []
198        for removed_node_id in self._unmatched_source_nodes:
199            edit_script.append(Remove(self._source_index[removed_node_id]))
200        for inserted_node_id in self._unmatched_target_nodes:
201            edit_script.append(Insert(self._target_index[inserted_node_id]))
202        for kept_source_node_id, kept_target_node_id in matching_set:
203            source_node = self._source_index[kept_source_node_id]
204            target_node = self._target_index[kept_target_node_id]
205
206            if (
207                not isinstance(source_node, UPDATABLE_EXPRESSION_TYPES)
208                or source_node == target_node
209            ):
210                edit_script.extend(
211                    self._generate_move_edits(source_node, target_node, matching_set)
212                )
213
214                source_non_expression_leaves = dict(_get_non_expression_leaves(source_node))
215                target_non_expression_leaves = dict(_get_non_expression_leaves(target_node))
216
217                if source_non_expression_leaves != target_non_expression_leaves:
218                    edit_script.append(Update(source_node, target_node))
219                elif not delta_only:
220                    edit_script.append(Keep(source_node, target_node))
221            else:
222                edit_script.append(Update(source_node, target_node))
223
224        return edit_script
225
226    def _generate_move_edits(
227        self, source: exp.Expression, target: exp.Expression, matching_set: t.Set[t.Tuple[int, int]]
228    ) -> t.List[Move]:
229        source_args = [id(e) for e in _expression_only_args(source)]
230        target_args = [id(e) for e in _expression_only_args(target)]
231
232        args_lcs = set(_lcs(source_args, target_args, lambda l, r: (l, r) in matching_set))
233
234        move_edits = []
235        for a in source_args:
236            if a not in args_lcs and a not in self._unmatched_source_nodes:
237                move_edits.append(Move(self._source_index[a]))
238
239        return move_edits
240
241    def _compute_matching_set(self) -> t.Set[t.Tuple[int, int]]:
242        leaves_matching_set = self._compute_leaf_matching_set()
243        matching_set = leaves_matching_set.copy()
244
245        ordered_unmatched_source_nodes = {
246            id(n): None for n in self._source.bfs() if id(n) in self._unmatched_source_nodes
247        }
248        ordered_unmatched_target_nodes = {
249            id(n): None for n in self._target.bfs() if id(n) in self._unmatched_target_nodes
250        }
251
252        for source_node_id in ordered_unmatched_source_nodes:
253            for target_node_id in ordered_unmatched_target_nodes:
254                source_node = self._source_index[source_node_id]
255                target_node = self._target_index[target_node_id]
256                if _is_same_type(source_node, target_node):
257                    source_leaf_ids = {id(l) for l in _get_expression_leaves(source_node)}
258                    target_leaf_ids = {id(l) for l in _get_expression_leaves(target_node)}
259
260                    max_leaves_num = max(len(source_leaf_ids), len(target_leaf_ids))
261                    if max_leaves_num:
262                        common_leaves_num = sum(
263                            1 if s in source_leaf_ids and t in target_leaf_ids else 0
264                            for s, t in leaves_matching_set
265                        )
266                        leaf_similarity_score = common_leaves_num / max_leaves_num
267                    else:
268                        leaf_similarity_score = 0.0
269
270                    adjusted_t = (
271                        self.t if min(len(source_leaf_ids), len(target_leaf_ids)) > 4 else 0.4
272                    )
273
274                    if leaf_similarity_score >= 0.8 or (
275                        leaf_similarity_score >= adjusted_t
276                        and self._dice_coefficient(source_node, target_node) >= self.f
277                    ):
278                        matching_set.add((source_node_id, target_node_id))
279                        self._unmatched_source_nodes.remove(source_node_id)
280                        self._unmatched_target_nodes.remove(target_node_id)
281                        ordered_unmatched_target_nodes.pop(target_node_id, None)
282                        break
283
284        return matching_set
285
286    def _compute_leaf_matching_set(self) -> t.Set[t.Tuple[int, int]]:
287        candidate_matchings: t.List[t.Tuple[float, int, int, exp.Expression, exp.Expression]] = []
288        source_expression_leaves = list(_get_expression_leaves(self._source))
289        target_expression_leaves = list(_get_expression_leaves(self._target))
290        for source_leaf in source_expression_leaves:
291            for target_leaf in target_expression_leaves:
292                if _is_same_type(source_leaf, target_leaf):
293                    similarity_score = self._dice_coefficient(source_leaf, target_leaf)
294                    if similarity_score >= self.f:
295                        heappush(
296                            candidate_matchings,
297                            (
298                                -similarity_score,
299                                -_parent_similarity_score(source_leaf, target_leaf),
300                                len(candidate_matchings),
301                                source_leaf,
302                                target_leaf,
303                            ),
304                        )
305
306        # Pick best matchings based on the highest score
307        matching_set = set()
308        while candidate_matchings:
309            _, _, _, source_leaf, target_leaf = heappop(candidate_matchings)
310            if (
311                id(source_leaf) in self._unmatched_source_nodes
312                and id(target_leaf) in self._unmatched_target_nodes
313            ):
314                matching_set.add((id(source_leaf), id(target_leaf)))
315                self._unmatched_source_nodes.remove(id(source_leaf))
316                self._unmatched_target_nodes.remove(id(target_leaf))
317
318        return matching_set
319
320    def _dice_coefficient(self, source: exp.Expression, target: exp.Expression) -> float:
321        source_histo = self._bigram_histo(source)
322        target_histo = self._bigram_histo(target)
323
324        total_grams = sum(source_histo.values()) + sum(target_histo.values())
325        if not total_grams:
326            return 1.0 if source == target else 0.0
327
328        overlap_len = 0
329        overlapping_grams = set(source_histo) & set(target_histo)
330        for g in overlapping_grams:
331            overlap_len += min(source_histo[g], target_histo[g])
332
333        return 2 * overlap_len / total_grams
334
335    def _bigram_histo(self, expression: exp.Expression) -> t.DefaultDict[str, int]:
336        if id(expression) in self._bigram_histo_cache:
337            return self._bigram_histo_cache[id(expression)]
338
339        expression_str = self._sql_generator.generate(expression)
340        count = max(0, len(expression_str) - 1)
341        bigram_histo: t.DefaultDict[str, int] = defaultdict(int)
342        for i in range(count):
343            bigram_histo[expression_str[i : i + 2]] += 1
344
345        self._bigram_histo_cache[id(expression)] = bigram_histo
346        return bigram_histo

The implementation of the Change Distiller algorithm described by Beat Fluri and Martin Pinzger in their paper https://ieeexplore.ieee.org/document/4339230, which in turn is based on the algorithm by Chawathe et al. described in http://ilpubs.stanford.edu:8090/115/1/1995-46.pdf.

ChangeDistiller( f: float = 0.6, t: float = 0.6, dialect: Union[str, sqlglot.dialects.dialect.Dialect, Type[sqlglot.dialects.dialect.Dialect], NoneType] = None)
160    def __init__(self, f: float = 0.6, t: float = 0.6, dialect: DialectType = None) -> None:
161        self.f = f
162        self.t = t
163        self._sql_generator = Dialect.get_or_raise(dialect).generator()
f
t
def diff( self, source: sqlglot.expressions.Expression, target: sqlglot.expressions.Expression, matchings: Optional[List[Tuple[sqlglot.expressions.Expression, sqlglot.expressions.Expression]]] = None, delta_only: bool = False) -> List[Union[Insert, Remove, Move, Update, Keep]]:
165    def diff(
166        self,
167        source: exp.Expression,
168        target: exp.Expression,
169        matchings: t.List[t.Tuple[exp.Expression, exp.Expression]] | None = None,
170        delta_only: bool = False,
171    ) -> t.List[Edit]:
172        matchings = matchings or []
173        pre_matched_nodes = {id(s): id(t) for s, t in matchings}
174        if len({n for pair in pre_matched_nodes.items() for n in pair}) != 2 * len(matchings):
175            raise ValueError("Each node can be referenced at most once in the list of matchings")
176
177        self._source = source
178        self._target = target
179        self._source_index = {
180            id(n): n for n in self._source.bfs() if not isinstance(n, IGNORED_LEAF_EXPRESSION_TYPES)
181        }
182        self._target_index = {
183            id(n): n for n in self._target.bfs() if not isinstance(n, IGNORED_LEAF_EXPRESSION_TYPES)
184        }
185        self._unmatched_source_nodes = set(self._source_index) - set(pre_matched_nodes)
186        self._unmatched_target_nodes = set(self._target_index) - set(pre_matched_nodes.values())
187        self._bigram_histo_cache: t.Dict[int, t.DefaultDict[str, int]] = {}
188
189        matching_set = self._compute_matching_set() | set(pre_matched_nodes.items())
190        return self._generate_edit_script(matching_set, delta_only)