Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

planner: derive index filters for mv index paths #54877

Merged
merged 4 commits into from
Jul 26, 2024
Merged

planner: derive index filters for mv index paths #54877

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
f672293
add
time-and-fate Jul 24, 2024
fa74f46
add test case and fix for prefix index case
time-and-fate Jul 24, 2024
4b4c19a
Merge remote-tracking branch 'upstream/master' into s24-mv-idx-split-…
time-and-fate Jul 24, 2024
053f681
Update pkg/planner/core/indexmerge_path.go
time-and-fate Jul 26, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
53 changes: 49 additions & 4 deletions pkg/planner/core/indexmerge_path.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -981,7 +981,29 @@ func (ds *DataSource) generateIndexMerge4ComposedIndex(normalPathCnt int, indexM
remainedCNFs = append(remainedCNFs, CNFItem)
}
}
mvp := ds.buildPartialPathUp4MVIndex(combinedPartialPaths, true, remainedCNFs, ds.TableStats.HistColl)

condInIdxFilter := make(map[string]struct{}, len(remainedCNFs))
// try to derive index filters for each path
for _, path := range combinedPartialPaths {
idxFilters, _ := ds.splitIndexFilterConditions(remainedCNFs, path.FullIdxCols, path.FullIdxColLens)
idxFilters = util.CloneExprs(idxFilters)
path.IndexFilters = append(path.IndexFilters, idxFilters...)
for _, idxFilter := range idxFilters {
condInIdxFilter[string(idxFilter.HashCode())] = struct{}{}
}
}

// Collect the table filters.
// Since it's the intersection type index merge here, as long as a filter appears in one path, we don't need it in
// the table filters.
var tableFilters []expression.Expression
for _, CNFItem := range remainedCNFs {
if _, ok := condInIdxFilter[string(CNFItem.HashCode())]; !ok {
tableFilters = append(tableFilters, CNFItem)
}
}

mvp := ds.buildPartialPathUp4MVIndex(combinedPartialPaths, true, tableFilters, ds.TableStats.HistColl)

ds.PossibleAccessPaths = append(ds.PossibleAccessPaths, mvp)
return nil
Expand Down Expand Up @@ -1046,10 +1068,27 @@ func (ds *DataSource) generateIndexMerge4MVIndex(normalPathCnt int, filters []ex
continue
}

// Here, all partial paths are built from the same MV index, so we can directly use the first one to get the
// metadata.
// And according to buildPartialPaths4MVIndex, there must be at least one partial path if it returns ok.
firstPath := partialPaths[0]
idxFilters, tableFilters := ds.splitIndexFilterConditions(
remainingFilters,
firstPath.FullIdxCols,
firstPath.FullIdxColLens,
)

// Add the index filters to every partial path.
// For union type index merge, this is necessary for correctness.
for _, path := range partialPaths {
clonedIdxFilters := util.CloneExprs(idxFilters)
path.IndexFilters = append(path.IndexFilters, clonedIdxFilters...)
}

ds.PossibleAccessPaths = append(ds.PossibleAccessPaths, ds.buildPartialPathUp4MVIndex(
partialPaths,
isIntersection,
remainingFilters,
tableFilters,
ds.TableStats.HistColl,
),
)
Expand Down Expand Up @@ -1214,10 +1253,16 @@ func buildPartialPath4MVIndex(
partialPath := &util.AccessPath{Index: mvIndex}
partialPath.Ranges = ranger.FullRange()
for i := 0; i < len(idxCols); i++ {
length := mvIndex.Columns[i].Length
// For full length prefix index, we consider it as non prefix index.
// This behavior is the same as in IndexInfo2Cols(), which is used for non mv index.
if length == idxCols[i].RetType.GetFlen() {
length = types.UnspecifiedLength
}
partialPath.IdxCols = append(partialPath.IdxCols, idxCols[i])
partialPath.IdxColLens = append(partialPath.IdxColLens, mvIndex.Columns[i].Length)
partialPath.IdxColLens = append(partialPath.IdxColLens, length)
partialPath.FullIdxCols = append(partialPath.FullIdxCols, idxCols[i])
partialPath.FullIdxColLens = append(partialPath.FullIdxColLens, mvIndex.Columns[i].Length)
partialPath.FullIdxColLens = append(partialPath.FullIdxColLens, length)
}
if err := detachCondAndBuildRangeForPath(sctx, partialPath, accessFilters, histColl); err != nil {
return nil, false, err
Expand Down
124 changes: 124 additions & 0 deletions tests/integrationtest/r/planner/core/indexmerge_path.result
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1081,3 +1081,127 @@ IndexMerge 0.05 root type: union
└─Selection(Probe) 0.05 cop[tikv] eq(planner__core__indexmerge_path.t.a, 1)
└─TableRowIDScan 49.94 cop[tikv] table:t keep order:false, stats:pseudo
SET @@tidb_opt_fix_control = default;
drop table if exists t, t1;
Copy link
Contributor

@AilinKid AilinKid Jul 26, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

the logic code change makes sense to me, by the way, is there any test case that needs to be modified?

create table t (a int, b varchar(30), c float, j json, pk int primary key,
key mvi1(c, (cast(j->'$.a' as unsigned array)), b),
key mvi2(a, (cast(j->'$.c' as unsigned array))),
key mvi3((cast(j->'$.d' as unsigned array)), c),
key idx(b, c)
);
insert into t values (1, 'test', 1, '{"a":[3,4,5], "c":[7,8,9], "d":[10,11,12]}', 1);
insert into t values (2, 'text', 1, '{"a":[4,5,6], "c":[10,11,12], "d":[13,14,15]}', 2);
insert into t values (1, 'abcd', 1, '{"a":[7,8,9], "c":[13,14,15], "d":[16,17,18]}', 3);
SELECT /*+ use_index_merge(t, mvi1) */ * from t where
c = 1 and
json_overlaps(j->'$.a', '[4,5,6]') and
b not like '%test%';
a b c j pk
2 text 1 {"a": [4, 5, 6], "c": [10, 11, 12], "d": [13, 14, 15]} 2
EXPLAIN format = brief SELECT /*+ use_index_merge(t, mvi1) */ * from t where
c = 1 and
json_overlaps(j->'$.a', '[4,5,6]') and
b not like '%test%';
id estRows task access object operator info
Selection 0.24 root json_overlaps(json_extract(planner__core__indexmerge_path.t.j, "$.a"), cast("[4,5,6]", json BINARY))
└─IndexMerge 0.30 root type: union
├─Selection(Build) 0.00 cop[tikv] not(like(planner__core__indexmerge_path.t.b, "%test%", 92))
│ └─IndexRangeScan 0.10 cop[tikv] table:t, index:mvi1(c, cast(json_extract(`j`, _utf8mb4'$.a') as unsigned array), b) range:[1 4,1 4], keep order:false, stats:pseudo
├─Selection(Build) 0.00 cop[tikv] not(like(planner__core__indexmerge_path.t.b, "%test%", 92))
│ └─IndexRangeScan 0.10 cop[tikv] table:t, index:mvi1(c, cast(json_extract(`j`, _utf8mb4'$.a') as unsigned array), b) range:[1 5,1 5], keep order:false, stats:pseudo
├─Selection(Build) 0.00 cop[tikv] not(like(planner__core__indexmerge_path.t.b, "%test%", 92))
│ └─IndexRangeScan 0.10 cop[tikv] table:t, index:mvi1(c, cast(json_extract(`j`, _utf8mb4'$.a') as unsigned array), b) range:[1 6,1 6], keep order:false, stats:pseudo
└─TableRowIDScan(Probe) 0.30 cop[tikv] table:t keep order:false, stats:pseudo
SELECT /*+ use_index_merge(t, mvi1) */ * from t where
c = 1 and
json_contains(j->'$.a', '[4,5]') and
b not like '%test%';
a b c j pk
2 text 1 {"a": [4, 5, 6], "c": [10, 11, 12], "d": [13, 14, 15]} 2
EXPLAIN format = brief SELECT /*+ use_index_merge(t, mvi1) */ * from t where
c = 1 and
json_contains(j->'$.a', '[4,5]') and
b not like '%test%';
id estRows task access object operator info
IndexMerge 0.00 root type: intersection
├─Selection(Build) 0.00 cop[tikv] not(like(planner__core__indexmerge_path.t.b, "%test%", 92))
│ └─IndexRangeScan 0.10 cop[tikv] table:t, index:mvi1(c, cast(json_extract(`j`, _utf8mb4'$.a') as unsigned array), b) range:[1 4,1 4], keep order:false, stats:pseudo
├─Selection(Build) 0.00 cop[tikv] not(like(planner__core__indexmerge_path.t.b, "%test%", 92))
│ └─IndexRangeScan 0.10 cop[tikv] table:t, index:mvi1(c, cast(json_extract(`j`, _utf8mb4'$.a') as unsigned array), b) range:[1 5,1 5], keep order:false, stats:pseudo
└─TableRowIDScan(Probe) 0.00 cop[tikv] table:t keep order:false, stats:pseudo
SELECT /*+ use_index_merge(t, mvi1, mvi2, idx) */ * from t where
a = 1 and
b > 'abc' and
b not like '%test%' and
c = 10 and
3 member of (j->'$.a') and
3 member of (j->'$.c');
a b c j pk
EXPLAIN format=brief SELECT /*+ use_index_merge(t, mvi1, mvi2, idx) */ * from t where
a = 1 and
b > 'abc' and
b not like '%test%' and
c = 10 and
3 member of (j->'$.a') and
3 member of (j->'$.c');
id estRows task access object operator info
IndexMerge 0.00 root type: intersection
├─Selection(Build) 2.67 cop[tikv] eq(planner__core__indexmerge_path.t.c, 10), not(like(planner__core__indexmerge_path.t.b, "%test%", 92)), not(like(planner__core__indexmerge_path.t.b, "%test%", 92))
│ └─IndexRangeScan 3333.33 cop[tikv] table:t, index:idx(b, c) range:("abc",+inf], keep order:false, stats:pseudo
├─Selection(Build) 0.00 cop[tikv] not(like(planner__core__indexmerge_path.t.b, "%test%", 92))
│ └─IndexRangeScan 0.10 cop[tikv] table:t, index:mvi1(c, cast(json_extract(`j`, _utf8mb4'$.a') as unsigned array), b) range:[10 3,10 3], keep order:false, stats:pseudo
├─IndexRangeScan(Build) 0.10 cop[tikv] table:t, index:mvi2(a, cast(json_extract(`j`, _utf8mb4'$.c') as unsigned array)) range:[1 3,1 3], keep order:false, stats:pseudo
└─TableRowIDScan(Probe) 0.00 cop[tikv] table:t keep order:false, stats:pseudo
create table t1 (
a int,
b varchar(30),
c float,
d tinytext,
j json,
key mvi1(b(3), (cast(j as unsigned array))),
key mvi2((cast(j as unsigned array)), b),
key mvi3((cast(j as unsigned array)), d(30)),
key mvi4((cast(j as unsigned array)), d(255))
);
EXPLAIN format = brief SELECT /*+ use_index_merge(t1, mvi1) */ * from t1 where
c = 1 and
json_contains(j, '[4,5]') and
b = 'abcdefg';
id estRows task access object operator info
TableReader 0.00 root data:Selection
└─Selection 0.00 cop[tikv] eq(planner__core__indexmerge_path.t1.b, "abcdefg"), eq(planner__core__indexmerge_path.t1.c, 1), json_contains(planner__core__indexmerge_path.t1.j, cast("[4,5]", json BINARY))
└─TableFullScan 10000.00 cop[tikv] table:t1 keep order:false, stats:pseudo
EXPLAIN format = brief SELECT /*+ use_index_merge(t1, mvi2) */ * from t1 where
c = 1 and
json_contains(j, '[4,5]') and
b = 'abcdefg' and
b like '%test%';
id estRows task access object operator info
IndexMerge 0.00 root type: intersection
├─Selection(Build) 0.00 cop[tikv] like(planner__core__indexmerge_path.t1.b, "%test%", 92)
│ └─IndexRangeScan 0.10 cop[tikv] table:t1, index:mvi2(cast(`j` as unsigned array), b) range:[4 "abcdefg",4 "abcdefg"], keep order:false, stats:pseudo
├─Selection(Build) 0.00 cop[tikv] like(planner__core__indexmerge_path.t1.b, "%test%", 92)
│ └─IndexRangeScan 0.10 cop[tikv] table:t1, index:mvi2(cast(`j` as unsigned array), b) range:[5 "abcdefg",5 "abcdefg"], keep order:false, stats:pseudo
└─Selection(Probe) 0.00 cop[tikv] eq(planner__core__indexmerge_path.t1.c, 1)
└─TableRowIDScan 0.00 cop[tikv] table:t1 keep order:false, stats:pseudo
EXPLAIN format = brief SELECT /*+ use_index_merge(t1, mvi3) */ * from t1 where
c = 1 and
json_contains(j, '[4,5]') and
d not like '%test%';
id estRows task access object operator info
IndexMerge 0.01 root type: intersection
├─IndexRangeScan(Build) 10.00 cop[tikv] table:t1, index:mvi3(cast(`j` as unsigned array), d) range:[4,4], keep order:false, stats:pseudo
├─IndexRangeScan(Build) 10.00 cop[tikv] table:t1, index:mvi3(cast(`j` as unsigned array), d) range:[5,5], keep order:false, stats:pseudo
└─Selection(Probe) 0.01 cop[tikv] eq(planner__core__indexmerge_path.t1.c, 1), not(like(planner__core__indexmerge_path.t1.d, "%test%", 92))
└─TableRowIDScan 0.01 cop[tikv] table:t1 keep order:false, stats:pseudo
EXPLAIN format = brief SELECT /*+ use_index_merge(t1, mvi4) */ * from t1 where
c = 1 and
json_contains(j, '[4,5]') and
d not like '%test%';
id estRows task access object operator info
IndexMerge 0.01 root type: intersection
├─Selection(Build) 0.00 cop[tikv] not(like(planner__core__indexmerge_path.t1.d, "%test%", 92))
│ └─IndexRangeScan 10.00 cop[tikv] table:t1, index:mvi4(cast(`j` as unsigned array), d) range:[4,4], keep order:false, stats:pseudo
├─Selection(Build) 0.00 cop[tikv] not(like(planner__core__indexmerge_path.t1.d, "%test%", 92))
│ └─IndexRangeScan 10.00 cop[tikv] table:t1, index:mvi4(cast(`j` as unsigned array), d) range:[5,5], keep order:false, stats:pseudo
└─Selection(Probe) 0.01 cop[tikv] eq(planner__core__indexmerge_path.t1.c, 1)
└─TableRowIDScan 0.01 cop[tikv] table:t1 keep order:false, stats:pseudo
88 changes: 88 additions & 0 deletions tests/integrationtest/t/planner/core/indexmerge_path.test
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -425,3 +425,91 @@ EXPLAIN format = brief SELECT * FROM t WHERE a > 1 AND (b = '2' OR c = 3 OR b =
EXPLAIN format = brief SELECT * FROM t WHERE a > 1 AND (b = '2' OR c = 3 OR b = '4' OR c = 5 OR b = '12' OR c = 13);
EXPLAIN format = brief SELECT * FROM t WHERE a = 1 AND (c = 13 OR c = 15 OR c = 5 OR b = '12' OR c = 13 OR b = '11');
SET @@tidb_opt_fix_control = default;

# Test deriving index filters for mv index paths
drop table if exists t, t1;
create table t (a int, b varchar(30), c float, j json, pk int primary key,
key mvi1(c, (cast(j->'$.a' as unsigned array)), b),
key mvi2(a, (cast(j->'$.c' as unsigned array))),
key mvi3((cast(j->'$.d' as unsigned array)), c),
key idx(b, c)
);

insert into t values (1, 'test', 1, '{"a":[3,4,5], "c":[7,8,9], "d":[10,11,12]}', 1);
insert into t values (2, 'text', 1, '{"a":[4,5,6], "c":[10,11,12], "d":[13,14,15]}', 2);
insert into t values (1, 'abcd', 1, '{"a":[7,8,9], "c":[13,14,15], "d":[16,17,18]}', 3);

# case 1: union type index merge on single mv index from single condition
SELECT /*+ use_index_merge(t, mvi1) */ * from t where
c = 1 and
json_overlaps(j->'$.a', '[4,5,6]') and
b not like '%test%';

EXPLAIN format = brief SELECT /*+ use_index_merge(t, mvi1) */ * from t where
c = 1 and
json_overlaps(j->'$.a', '[4,5,6]') and
b not like '%test%';

# case 2: intersection type index merge on single mv index from single condition
SELECT /*+ use_index_merge(t, mvi1) */ * from t where
c = 1 and
json_contains(j->'$.a', '[4,5]') and
b not like '%test%';

EXPLAIN format = brief SELECT /*+ use_index_merge(t, mvi1) */ * from t where
c = 1 and
json_contains(j->'$.a', '[4,5]') and
b not like '%test%';

# case 3: intersection type index merge on multiple indexes from different conditions
SELECT /*+ use_index_merge(t, mvi1, mvi2, idx) */ * from t where
a = 1 and
b > 'abc' and
b not like '%test%' and
c = 10 and
3 member of (j->'$.a') and
3 member of (j->'$.c');

EXPLAIN format=brief SELECT /*+ use_index_merge(t, mvi1, mvi2, idx) */ * from t where
a = 1 and
b > 'abc' and
b not like '%test%' and
c = 10 and
3 member of (j->'$.a') and
3 member of (j->'$.c');

# case 4: test prefix indexes
create table t1 (
a int,
b varchar(30),
c float,
d tinytext,
j json,
key mvi1(b(3), (cast(j as unsigned array))),
key mvi2((cast(j as unsigned array)), b),
key mvi3((cast(j as unsigned array)), d(30)),
key mvi4((cast(j as unsigned array)), d(255))
);


EXPLAIN format = brief SELECT /*+ use_index_merge(t1, mvi1) */ * from t1 where
c = 1 and
json_contains(j, '[4,5]') and
b = 'abcdefg';

EXPLAIN format = brief SELECT /*+ use_index_merge(t1, mvi2) */ * from t1 where
c = 1 and
json_contains(j, '[4,5]') and
b = 'abcdefg' and
b like '%test%';

EXPLAIN format = brief SELECT /*+ use_index_merge(t1, mvi3) */ * from t1 where
c = 1 and
json_contains(j, '[4,5]') and
d not like '%test%';

EXPLAIN format = brief SELECT /*+ use_index_merge(t1, mvi4) */ * from t1 where
c = 1 and
json_contains(j, '[4,5]') and
d not like '%test%';