Re: [PATCH] Use Boyer-Moore-Horspool for simple LIKE contains patterns

On Wed, Sep 16, 2026 at 8:37 AM Atsushi Ogawa
<atsushi(dot)ogawa001(at)gmail(dot)com> wrote:
>
>
> Hi Haibo,
>
> Thanks for the detailed matrix. I reproduced the low-entropy regression
> region and agree with your reading: the cost is driven by alphabet size
> and mismatch position rather than literal length, so adjusting
> LIKE_BMH_MIN_LITERAL_LEN cannot describe the boundary.
>
> I would like to get your thoughts on the general approach first.
> I've attached a rough PoC patch just for reference; it still needs
> some cleanup, and a proper patch along with the full numbers
> will follow later.
>
> Needle-only rule
> ----------------
>
> I first tried a needle-only check: skip BMH when the trailing bytes of the
> literal are periodic. While it handles cases like repeat('a') and certain
> repeat('ab') patterns, it does nothing when the periodicity lies in the
> haystack rather than the literal (e.g., repeat('abcd') with a 64-byte
> literal stays around 15x slower). It also needlessly forces literals like
> '%aaaa%' to the generic matcher on ordinary text where BMH would otherwise
> win. A preparation-time test on the pattern alone does not seem viable.
>
> Bounded work with resume
> ------------------------
>
> The approach I am leaning towards is a runtime guard along the lines of
> your suggestion, structured as follows:
>
> - like_bmh_search() checks the guard (last) byte first and only counts
> inner-loop byte comparisons beyond that guard. The fast path where the
> last byte differs has zero accounting overhead, preserving standard BMH
> performance.
>
> - When the comparison count exceeds a given threshold (currently
> prototyping slen / 2), BMH aborts and reports the offset up to which it
> has ruled out matches.
>
> - LikeMatchText() then delegates only the remaining, unsearched suffix
> (backed up to a character boundary, tested under UTF-8) to GenericMatchText().
> Because the pattern begins with '%', evaluating the suffix yields the
> exact same semantics without rescanning the entire string from the
> start.
>
> Preliminary numbers (100,000 rows, best of 5, ms; HEAD / v3 / v3 + bounded work):
>
> repeat('a',1024) LIKE '%~aaaaaaaaaaaaaaa%' 111 / 552 / 131
> repeat('a',1024) LIKE '%~' || 63 x 'a' || '%' 113 / 2007 / 135
> repeat('a',1024) LIKE '%aaaaaaaaaaaaaaa~%' 1896 / 353 / 325
> English text LIKE '%worst of crimes%' (miss) 162 / 65 / 47
>
> In a microbenchmark, the worst-case late-mismatch penalty drops from
> ~100x down to ~4x at 1024 bytes (~2.3x at 32 bytes). Early mismatch and
> match-present cases remain unaffected or slightly faster thanks to the
> guard byte.
>
> A residual 1.2-4x overhead remains in cases where the generic matcher
> quickly bails out (e.g., the first pattern byte is absent from the haystack)
> while BMH exhausts its comparison budget before falling back.
> Tightening the budget lowers this ceiling, but also trims BMH's advantages on
> benign inputs.
>
> As a note, the attached PoC is an incremental patch on top of v3 rather
> than HEAD.
>
> Regards,
> Atsushi Ogawa

Hi Atsushi,

I took a closer look at the PoC and tested the bounded-work/resume path fairly
aggressively. The general approach looks sound to me.

In particular, I was able to convince myself that the resume offset is correct.
With

searched = pos - literal_len + 2

the returned position is the earliest start that has not already been ruled out
by the completed Horspool alignment. The budget check happens only after the
current candidate comparison has completed, so there is no partially examined
alignment to account for. Backing up to a UTF-8 character boundary only enlarges
the suffix passed to GenericMatchText(), which is safe for %literal%.

I also ran exhaustive/fuzz differential tests around the resume boundary,
including UTF-8 and single-byte cases, and did not find a mismatch.

The performance results are encouraging as well. On the previous matrix, using
the current slen / 2 budget changed the overall result roughly as follows:

v3 unbounded bounded
median ratio 0.314x 0.314x
worst regression 33.30x 1.80x
cases > 1.25x 98 64
cases > 2x 62 0
>= 1.25x wins retained 89.4%
>= 2x wins retained 87.8%

So the >2x regression region disappears while most of the useful BMH
wins remain.

I did find one issue with the current budget definition. If

slen / 2 < literal_len

then the first candidate can consume the entire budget, after which
GenericMatchText() rescans almost the whole haystack. In my matrix this turned
seven cases from roughly 0.46x wins into 1.46-1.49x regressions.

A simple lower bound seems to avoid that class:

budget = Max(slen / 2, literal_len)

More interestingly, I also tried several budget values, and

Max(slen / 4, literal_len)

looked better than slen / 2 in this test set. It kept essentially the same
median, worst case, and useful-win retention, but reduced the number of >1.25x
regressions from 64 to 24 and improved the p95 ratio from about 1.31x to 1.23x.

I would not read too much into /4 as a magic constant yet, but I think the
literal_len lower bound is important, and /4 seems worth including in the full
benchmark when you prepare the next patch.

One other thing I found is that the remaining ~1.7-1.8x matcher-level
regressions do not appear to come from exhausting the budget too late. In some
of those cases BMH performs fewer byte comparisons than GenericMatchText(),
but still loses because it executes many dependent skip-table lookups with a
small average skip. Tightening the budget further starts to remove substantial
BMH wins, so I don’t think that residual can be eliminated cleanly with a
smaller threshold alone.

I also checked the “zero accounting overhead” point. On the guard-miss path
the bounded version generates essentially the same fast path as a guard-first
version without accounting; the improvement over v3 appears to come from the
guard-first restructuring itself rather than measurement noise.

So my current view is:

1. the bounded-work + resume design looks correct;
2. it addresses the serious regression region very effectively;
3. the budget should probably have a lower bound of literal_len;
4. Max(slen / 4, literal_len) looks worth testing alongside /2;
5. beyond that, the remaining small regression region looks more like an
inherent BMH cost than a failure of the fallback policy.

Thanks,
Haibo

>
>
> 2026年9月15日(火) 10:24 Haibo Yan <tristan(dot)yim(at)gmail(dot)com>:
>>
>> On Fri, Jul 17, 2026 at 3:23 AM Atsushi Ogawa
>> <atsushi(dot)ogawa001(at)gmail(dot)com> wrote:
>> >
>> > Hi Greg,
>> >
>> > Thanks for the careful review. I have attached a v2 patch.
>> >
>> > > git grep shows we already use BMH in src/backend/utils/adt/varlena.c
>> > > Worth acknowledging that in a code comment somewhere? I didn't see any
>> > > obvious advantage to refactoring things out at quick glance, but a mention
>> > > might be nice.
>> >
>> > Agreed. I added a comment at the top of like_bmh.c that cross-references the
>> > existing Boyer-Moore-Horspool implementation in varlena.c and explains why I
>> > kept the implementations separate. The varlena.c code searches one
>> > (haystack, needle) pair with an adaptively sized skip table, whereas the LIKE
>> > path interprets its internal backslash escapes while extracting the literal
>> > and caches the prepared search state in FmgrInfo for use across rows. I did
>> > not find a clean way to share that machinery without introducing more coupling
>> > than seemed useful.
>> >
>> > > + * by '%' wildcards. Remove backslash escapes while building the search
>> > > + * state.
>> > >
>> > > Slightly off comment. This is for like_bmh_pattern_is_eligible - we are not
>> > > removing here, just skipping things when we count.
>> >
>> > Right. I reworded the comment to say that the eligibility check skips
>> > backslash escapes while counting the literal length. The escapes are removed
>> > later, when the search state is built.
>> >
>> > > if (i + 1 >= plen - 1)
>> > >
>> > > Worth a comment to explain that we are catching the '%foo\%' case here.
>> >
>> > Added. The new comment explains that this rejects patterns such as
>> > '%foo\%', where the backslash escapes the closing '%' rather than a literal
>> > byte.
>> >
>> > > pattern_stable = get_fn_expr_arg_stable(flinfo, 1);
>> > >
>> > > /*
>> > > * ScalarArrayOpExpr invokes the operator once per array element. The
>> > > * array expression can be stable while the pattern passed to this function
>> > > * changes between calls, so it must not use a cached search state.
>> > > */
>> > > if (flinfo->fn_expr != NULL && IsA(flinfo->fn_expr, ScalarArrayOpExpr))
>> > > pattern_stable = false;
>> > >
>> > > My first thought was to make this an if/else so we don't reclobber, but
>> > > seeing how later on we check collation every time, I'm wondering if we
>> > > shouldn't just check the pattern as well every time via a memcmp like
>> > > regexp.c does in RE_compile_and_cache (and remove that block above).
>> > > So we store it verbatim in the like_bmh_init() function with memcpy, then
>> > > make the check inside like_bmh_match() that looks like this:
>> > >
>> > > unlikely(collation has changed)
>> > >
>> > > into:
>> > >
>> > > unlikely(
>> > > collation has changed
>> > > OR pattern length has changed
>> > > OR pattern itself has changed (e.g. memcmp true)
>> > > )
>> > >
>> > > Also means you could then roll get_fn_expr_arg_stable into that big old ||
>> > > grouping, and remove pattern_stable entirely.
>> >
>> > I implemented the suggested verbatim-pattern cache and benchmarked it directly
>> > against the initial patch's structural-stability design. The test scanned two
>> > million rows per transaction, with a warmup followed by the median of seven
>> > pgbench runs of 40 transactions each. The benchmark used an AMD EPYC 7763
>> > host with 8 vCPUs, GCC 11.4.0, and an -O2 -g build, using a UTF-8 database
>> > with C locale. The results below are median latency per scan:
>> >
>> > case initial patch memcmp vs. initial
>> > ------------------------------------ ------------- -------- -----------
>> > constant, 4-byte literal 63.7 ms 65.1 ms +2.3%
>> > constant, 32-byte literal 48.8 ms 48.4 ms -0.8%
>> > constant, 4-byte literal, 8-byte input 43.2 ms 44.0 ms +1.8%
>> > non-constant, fixed value at runtime 92.9 ms 57.0 ms -38.6%
>> > non-constant, changes on every row 91.6 ms 172.4 ms +88.2%
>> >
>> > The per-row length check and memcmp were therefore not particularly expensive
>> > for stable constant patterns. The more important tradeoff involved
>> > non-constant patterns. When the value remained fixed at runtime, the verbatim
>> > cache was faster because it could use BMH. When the pattern changed on every
>> > row, however, it was substantially slower than the initial patch, which sends
>> > that case to the existing generic matcher. The verbatim variant had to repeat
>> > the eligibility check and rebuild the 256-entry skip table for every row.
>> >
>> > I then tested a hybrid of the two approaches. Patterns that
>> > get_fn_expr_arg_stable() identifies as a Const or external Param keep the
>> > existing comparison-free search state. An eligible non-stable pattern stores
>> > its verbatim bytes and is revalidated with a length check and memcmp. On the
>> > first mismatch, the state is changed permanently to the generic marker. The
>> > mismatching row and all later rows use the existing matcher; the eligibility
>> > check and skip-table build are never repeated.
>> >
>> > ScalarArrayOpExpr still has to be classified as non-stable, since its array
>> > expression can be a Const while the operator receives a different element on
>> > each call. It now uses the same revalidation path and falls back permanently
>> > if the elements differ.
>> >
>> > I reran the comparison on aarch64 using two clean build trees based on the
>> > same source revision and configured with the same options. Both servers used
>> > the same data directory. The table contained two million 32-byte strings, a
>> > fixed pattern column, and an alternating pattern column. Parallel query was
>> > disabled, each server was warmed before measurement, and the server order was
>> > alternated in ABBA order. The figures below are medians of 16 EXPLAIN
>> > (ANALYZE, TIMING OFF) runs:
>> >
>> > case initial patch hybrid vs. initial
>> > -------------------------------- ------------- -------- -----------
>> > constant pattern 194.1 ms 185.6 ms -4.4%
>> > non-constant, fixed at runtime 299.8 ms 199.9 ms -33.3%
>> > non-constant, changes every row 303.4 ms 304.9 ms +0.5%
>> > generic fallback control 288.5 ms 289.5 ms +0.3%
>> >
>> > The constant-pattern difference appears to be a compiler-dependent code-layout
>> > effect rather than a benefit of the hybrid design, so I do not interpret it as
>> > a general speedup. More importantly, the runtime-fixed case captures the
>> > benefit of the verbatim cache, while the row-varying case tracks the generic
>> > fallback control instead of rebuilding the 256-entry skip table for every row.
>> >
>> > The attached v2 patch uses this hybrid design. Thus the common stable path
>> > does not pay a memcmp, runtime-fixed non-constant values can use BMH, and a
>> > pattern that is observed to vary falls back without any rebuild penalty.
>> >
>> > > Hm...that collation test and message is already caught and done by
>> > > GenericMatchText, so you could throw !OidIsValid(collation) into that ||
>> > > group as well, and remove the ereport section entirely. It then falls
>> > > through later to GenericMatchText, which complains about the collation
>> > > there.
>> >
>> > Done. The invalid-collation case is now included in the rejection group and
>> > falls through to GenericMatchText. I removed the duplicate ereport block from
>> > like_bmh.c.
>> >
>> > > It did have one test failure:
>> > >
>> > > @@ -151,8 +151,8 @@
>> > > p | matched
>> > > --------+---------
>> > > %abcd% | t
>> > > - %b%e% | f
>> > > %b_d% | t
>> > > + %b%e% | f
>> > > %wxyz% | f
>> > > (4 rows)
>> > >
>> > > I think it's from the "Row-varying patterns must use the generic matcher."
>> > > test.
>> >
>> > Thanks for catching this. This was a locale-dependent sort-order issue in the
>> > test, not a matcher failure. The query now uses ORDER BY p COLLATE "C".
>> >
>> > I retested the revised patch against PostgreSQL HEAD 0348090: all 246 core
>> > regression tests passed, including like_bmh, and all four contrib/pg_trgm
>> > tests passed.
>> >
>> > Thanks,
>> > Atsushi Ogawa
>> >
>> > 2026年7月15日(水) 3:29 Greg Sabino Mullane <htamfids(at)gmail(dot)com>:
>> >>
>> >> Great idea, love seeing the speedups! Also appreciate the background, detailed explanation, and benchmarks. Quick code review:
>> >>
>> >> git grep shows we already use BMH in src/backend/utils/adt/varlena.c
>> >> Worth acknowledging that in a code comment somewhere? I didn't see any obvious advantage to refactoring things out at quick glance, but a mention might be nice.
>> >>
>> >> > + * by '%' wildcards. Remove backslash escapes while building the search state.
>> >>
>> >> Slightly off comment. This is for like_bmh_pattern_is_eligible - we are not removing here, just skipping things when we count.
>> >>
>> >> > if (i + 1 >= plen - 1)
>> >>
>> >> Worth a comment to explain that we are catching the '%foo\%' case here.
>> >>
>> >>
>> >> > pattern_stable = get_fn_expr_arg_stable(flinfo, 1);
>> >> >
>> >> > /*
>> >> > * ScalarArrayOpExpr invokes the operator once per array element. The
>> >> > * array expression can be stable while the pattern passed to this function
>> >> > * changes between calls, so it must not use a cached search state.
>> >> > */
>> >> > if (flinfo->fn_expr != NULL && IsA(flinfo->fn_expr, ScalarArrayOpExpr))
>> >> > pattern_stable = false;
>> >>
>> >> My first thought was to make this an if/else so we don't reclobber, but seeing how later on we check collation every time, I'm wondering if we shouldn't just check the pattern as well every time via a memcmp like regexp.c does in RE_compile_and_cache (and remove that block above). So we store it verbatim in the like_bmh_init() function with memcpy, then make the check inside like_bmh_match() that looks like this:
>> >>
>> >> unlikely(collation has changed)
>> >>
>> >> into:
>> >>
>> >> unlikely(
>> >> collation has changed
>> >> OR pattern length has changed
>> >> OR pattern itself has changed (e.g. memcmp true)
>> >> )
>> >>
>> >> Also means you could then roll get_fn_expr_arg_stable into that big old || grouping, and remove pattern_stable entirely.
>> >>
>> >> Hm...that collation test and message is already caught and done by GenericMatchText, so you could throw !OidIsValid(collation) into that || group as well, and remove the ereport section entirely. It then falls through later to GenericMatchText, which complains about the collation there.
>> >>
>> >> Anyway, the patch compiled cleanly against d15a6bc2 (Tue Jul 14 10:28:04 2026 +0200)
>> >>
>> >> It did have one test failure:
>> >>
>> >> @@ -151,8 +151,8 @@
>> >> p | matched
>> >> --------+---------
>> >> %abcd% | t
>> >> - %b%e% | f
>> >> %b_d% | t
>> >> + %b%e% | f
>> >> %wxyz% | f
>> >> (4 rows)
>> >>
>> >> I think it's from the "Row-varying patterns must use the generic matcher." test.
>> >>
>> >>
>> >> Cheers,
>> >> Greg
>> >>
>>
>> Hi Ogawa-san,
>>
>> I did some more testing of the BMH fast path, specifically to check whether the
>> repetitive-input regression I mentioned is just one adversarial construction or
>> part of a broader pattern.
>>
>> I ran a matrix varying haystack structure, literal length, mismatch position,
>> and haystack length, with the existing LIKE matcher and the patch's BMH search
>> in the same binary. The overall result is actually quite favorable to BMH: it
>> wins most of the tested cases, often by a large margin. However, there is also a
>> fairly well-defined regression region on low-entropy inputs when the backwards
>> comparison fails late.
>>
>> A few representative numbers are:
>>
>> haystack literal length existing LIKE
>> BMH ratio
>> repeat('a', 1024) 16 2.14 ms
>> 20.78 ms 9.7x
>> repeat('a', 1024) 64 2.13 ms
>> 66.93 ms 31.4x
>> repeat('ab', ...), 1024 64 2.12 ms
>> 33.24 ms 15.7x
>> repeat('abcd', ...), 1024 64 2.12 ms
>> 16.70 ms 7.9x
>> random alphabet=4, 1024 4 2.24 ms 7.41 ms 3.3x
>> English text, 1024 16 2.23 ms
>> 1.40 ms 0.63x
>>
>> The important part seems to be the combination of low effective alphabet size
>> and mismatch position, rather than literal length by itself.
>>
>> For example, against `repeat('a', 1024)`, a literal shaped roughly as
>>
>> ~aaaaaaaaaaaaaaa
>>
>> causes Horspool to compare almost the whole literal backwards before failing,
>> while the skip for `a` is only one byte. For a 16-byte literal I counted 1009
>> candidate alignments and 16 comparisons per alignment. The existing LIKE matcher
>> has almost the opposite behavior here: its first literal byte (`~`) is absent
>> from the haystack, so it rejects candidates very cheaply.
>>
>> Mismatch position changes the result dramatically. With the same 1024-byte
>> repeated-`a` input and a 16-byte literal I measured approximately:
>>
>> immediate mismatch: BMH / existing LIKE = 0.04x
>> middle mismatch: = 0.34x
>> late mismatch: = 9.7x
>>
>> So BMH can be much faster or much slower on very similar inputs.
>>
>> This also means that increasing `LIKE_BMH_MIN_LITERAL_LEN` does not appear to
>> address the issue. The worst measured regression actually increased with literal
>> length:
>>
>> 4 bytes 4.3x
>> 8 bytes 5.1x
>> 16 bytes 9.9x
>> 32 bytes 18.5x
>> 64 bytes 33.7x
>>
>> The regression is not universal. In this test set, English text and random data
>> over medium/large alphabets did not show >2x regressions, and for literals >= 8
>> bytes they did not show meaningful regressions at all. So I would describe this
>> as a narrow but systematic low-entropy case rather than a general
>> problem with BMH.
>>
>> I also tried looking for a cheap needle-only rule that could avoid the
>> bad cases.
>> There are some useful signals in the skip table, but they have substantial false
>> positives. More fundamentally, the same byte-identical literal can be a large
>> win or a large loss depending only on the haystack/match position, so a
>> preparation-time test based only on the literal cannot completely solve this.
>>
>> This seems related to the concerns raised in the earlier BMH/LIKE discussions:
>>
>> https://www.postgresql.org/message-id/CALkFZpcbipVJO%3DxVvNQMZ7uLUgHzBn65GdjtBHdeb47QV4XzLw%40mail.gmail.com
>>
>> and Tom Lane's later discussion here:
>>
>> https://www.postgresql.org/message-id/3811203.1675907383%40sss.pgh.pa.us
>>
>> There is also the recent related thread here:
>>
>> https://www.postgresql.org/message-id/flat/88272f23-19b4-493d-bdd7-258218b74881%40gmail.com
>>
>> Given that this is a performance optimization, I think it would be useful to
>> decide explicitly how much regression on this class of inputs is acceptable, or
>> whether some bounded-work fallback would make sense. A runtime guard might be
>> more promising than a needle-only eligibility rule, since it could notice that
>> the search is doing unusually large amounts of work without requiring a separate
>> scan of the haystack.
>>
>> I don't think these results argue against using BMH in general — in the same
>> matrix it was substantially faster in most cases — but they do suggest that
>> the current literal-length threshold alone doesn't describe the profitability
>> boundary.
>>
>> Regards,
>> Haibo