Python: Minor cleanup for string pool interaction#14591
Merged
RasmusWL merged 1 commit intogithub:mainfrom Nov 6, 2023
Merged
Conversation
This doesn't really seem to impact performance, is just a bit cleaner way to write the QL code.
In the worst case, since we do string concat with empty string, we would need to load string + attempt to store new string in string pool, which is useless since we can surely just reuse the integer ID.
-- Previously
Evaluated non-recursive predicate ImportResolution#f829f0f0::ImportResolution::getImmediateModuleReference#1#ff#reorder_1_0#join_rhs#CSE@5fdc6eia in 1ms (size: 2).
Evaluated relational algebra for predicate ImportResolution#f829f0f0::ImportResolution::getImmediateModuleReference#1#ff#reorder_1_0#join_rhs#CSE@5fdc6eia with tuple counts:
1 ~0% {1} r1 = CONSTANT(unique string)[""]
1 ~0% {1} r2 = CONSTANT(unique string)[".__init__"]
2 ~0% {1} r3 = r1 UNION r2
return r3
Evaluated recursive predicate ImportResolution#f829f0f0::ImportResolution::getImmediateModuleReference#1#ff#reorder_1_0@b4f54xlv in 72ms on iteration 1 (delta size: 26339).
Evaluated relational algebra for predicate ImportResolution#f829f0f0::ImportResolution::getImmediateModuleReference#1#ff#reorder_1_0@b4f54xlv on iteration 1 running pipeline base with tuple counts:
6382 ~0% {3} r1 = JOIN ImportResolution#f829f0f0::ImportResolution::getImmediateModuleReference#1#ff#reorder_1_0#join_rhs#CSE WITH Module#68efc075::Module::getFile#0#dispred#ff CARTESIAN PRODUCT OUTPUT Rhs.1, Lhs.0, Rhs.0
6066 ~2% {4} r2 = JOIN r1 WITH Module#68efc075::isPreferredModuleForName#2#ff ON FIRST 1 OUTPUT Lhs.2, _, Lhs.1, Rhs.1
{2} r3 = REWRITE r2 WITH Tmp.1 := "", Out.1 := InverseAppend(Tmp.1,In.2,In.3) KEEPING 2
3721 ~0% {2} r4 = SCAN r3 OUTPUT In.1, In.0
25945 ~2% {2} r5 = JOIN r4 WITH ImportResolution#f829f0f0::ImportResolution::getReferenceToModuleName#1#ff ON FIRST 1 OUTPUT Rhs.1, Lhs.1
38 ~4% {3} r6 = SCAN SsaDefinitions#9197156e::SsaSource::init_module_submodule_defn#2#ff OUTPUT In.1, _, In.0
38 ~1% {3} r7 = REWRITE r6 WITH Out.1 := -1
38 ~1% {2} r8 = JOIN r7 WITH py_scope_flow_021#join_rhs ON FIRST 2 OUTPUT Rhs.2, Lhs.2
38 ~0% {2} r9 = JOIN r8 WITH Module#68efc075::Module::getPackageName#0#dispred#ff ON FIRST 1 OUTPUT Lhs.1, Rhs.1
394 ~2% {2} r10 = JOIN r9 WITH Essa#24e22a14::EssaVariable::getSourceVariable#0#dispred#ff_10#join_rhs ON FIRST 1 OUTPUT Rhs.1, Lhs.1
394 ~0% {3} r11 = JOIN r10 WITH DataFlowPublic#c4b6ce7e::EssaNode#ff_10#join_rhs ON FIRST 1 OUTPUT Lhs.0, Lhs.1, Rhs.1
394 ~0% {5} r12 = JOIN r11 WITH Essa#24e22a14::EssaVariable::getName#0#dispred#ff ON FIRST 1 OUTPUT _, Lhs.2, Lhs.1, Rhs.1, _
394 ~0% {2} r13 = REWRITE r12 WITH Tmp.0 := ".", Tmp.4 := "", Out.0 := (In.2 ++ Tmp.0 ++ In.3 ++ Tmp.4) KEEPING 2
394 ~0% {5} r14 = JOIN r11 WITH Essa#24e22a14::EssaVariable::getName#0#dispred#ff ON FIRST 1 OUTPUT _, Lhs.2, Lhs.1, Rhs.1, _
394 ~1% {2} r15 = REWRITE r14 WITH Tmp.0 := ".", Tmp.4 := ".__init__", Out.0 := (In.2 ++ Tmp.0 ++ In.3 ++ Tmp.4) KEEPING 2
788 ~3% {2} r16 = r13 UNION r15
401 ~2% {2} r17 = JOIN r16 WITH Module#68efc075::isPreferredModuleForName#2#ff_10#join_rhs ON FIRST 1 OUTPUT Rhs.1, Lhs.1
394 ~1% {2} r18 = JOIN r17 WITH Module#68efc075::Module::getFile#0#dispred#ff_10#join_rhs ON FIRST 1 OUTPUT Lhs.1, Rhs.1
26339 ~1% {2} r19 = r5 UNION r18
return r19
-- Now
Evaluated recursive predicate ImportResolution#f829f0f0::ImportResolution::getImmediateModuleReference#1#ff#reorder_1_0@409a0xhe in 8ms on iteration 1 (delta size: 26339).
Evaluated relational algebra for predicate ImportResolution#f829f0f0::ImportResolution::getImmediateModuleReference#1#ff#reorder_1_0@409a0xhe on iteration 1 running pipeline base with tuple counts:
3033 ~0% {2} r1 = JOIN Module#68efc075::Module::getFile#0#dispred#ff_10#join_rhs WITH Module#68efc075::isPreferredModuleForName#2#ff ON FIRST 1 OUTPUT Rhs.1, Lhs.1
3033 ~0% {4} r2 = JOIN Module#68efc075::Module::getFile#0#dispred#ff_10#join_rhs WITH Module#68efc075::isPreferredModuleForName#2#ff ON FIRST 1 OUTPUT Lhs.1, _, Rhs.1, _
{2} r3 = REWRITE r2 WITH Tmp.1 := "", Tmp.3 := ".__init__", Out.1 := InverseAppend(Tmp.1,Tmp.3,In.2) KEEPING 2
688 ~0% {2} r4 = SCAN r3 OUTPUT In.1, In.0
3721 ~0% {2} r5 = r1 UNION r4
25945 ~2% {2} r6 = JOIN r5 WITH ImportResolution#f829f0f0::ImportResolution::getReferenceToModuleName#1#ff ON FIRST 1 OUTPUT Rhs.1, Lhs.1
38 ~4% {3} r7 = SCAN SsaDefinitions#9197156e::SsaSource::init_module_submodule_defn#2#ff OUTPUT In.1, _, In.0
38 ~1% {3} r8 = REWRITE r7 WITH Out.1 := -1
38 ~1% {2} r9 = JOIN r8 WITH py_scope_flow_021#join_rhs ON FIRST 2 OUTPUT Rhs.2, Lhs.2
38 ~0% {2} r10 = JOIN r9 WITH Module#68efc075::Module::getPackageName#0#dispred#ff ON FIRST 1 OUTPUT Lhs.1, Rhs.1
394 ~2% {2} r11 = JOIN r10 WITH Essa#24e22a14::EssaVariable::getSourceVariable#0#dispred#ff_10#join_rhs ON FIRST 1 OUTPUT Rhs.1, Lhs.1
394 ~0% {3} r12 = JOIN r11 WITH DataFlowPublic#c4b6ce7e::EssaNode#ff_10#join_rhs ON FIRST 1 OUTPUT Lhs.0, Lhs.1, Rhs.1
394 ~2% {4} r13 = JOIN r12 WITH Essa#24e22a14::EssaVariable::getName#0#dispred#ff ON FIRST 1 OUTPUT _, Lhs.2, Lhs.1, Rhs.1
394 ~0% {2} r14 = REWRITE r13 WITH Tmp.0 := ".", Out.0 := (In.2 ++ Tmp.0 ++ In.3) KEEPING 2
394 ~0% {5} r15 = JOIN r12 WITH Essa#24e22a14::EssaVariable::getName#0#dispred#ff ON FIRST 1 OUTPUT _, Lhs.2, Lhs.1, Rhs.1, _
394 ~5% {2} r16 = REWRITE r15 WITH Tmp.0 := ".", Tmp.0 := (In.2 ++ Tmp.0 ++ In.3), Tmp.4 := ".__init__", Out.0 := (Tmp.0 ++ Tmp.4) KEEPING 2
788 ~0% {2} r17 = r14 UNION r16
401 ~1% {2} r18 = JOIN r17 WITH Module#68efc075::isPreferredModuleForName#2#ff_10#join_rhs ON FIRST 1 OUTPUT Rhs.1, Lhs.1
394 ~1% {2} r19 = JOIN r18 WITH Module#68efc075::Module::getFile#0#dispred#ff_10#join_rhs ON FIRST 1 OUTPUT Lhs.1, Rhs.1
26339 ~0% {2} r20 = r6 UNION r19
return r20
tausbn
approved these changes
Oct 25, 2023
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This doesn't really seem to impact performance, is just a bit cleaner way to write the QL code.
In the worst case, since we do string concat with empty string, we would need to load string + attempt to store new string in string pool, which is useless since we can surely just reuse the integer ID.
Previously
Now