Files
Vitaly PryakhinandClaude Opus 5.5 b0ad5072d5 Alias invalid UTF-8 queries in to_valid_utf8 perf test
The raw bytes in the column name broke `clickhouse_driver`.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-13 00:18:01 +03:00

63 lines
1.9 KiB
XML

<test>
<settings>
<max_threads>1</max_threads>
</settings>
<!-- Keep this to ten explicit cases: CI samples at most ten queries from each performance test. -->
<query>
SELECT sum(length(toValidUTF8(materialize(repeat('a', 127)))))
FROM numbers(100000)
FORMAT Null
</query>
<query>
SELECT sum(length(toValidUTF8(materialize(repeat('a', 128)))))
FROM numbers(100000)
FORMAT Null
</query>
<query>
SELECT sum(length(toValidUTF8(materialize(repeat('a', 129)))))
FROM numbers(100000)
FORMAT Null
</query>
<query>
SELECT sum(length(toValidUTF8(materialize(concat(repeat('é', 63), 'a')))))
FROM numbers(100000)
FORMAT Null
</query>
<query>
SELECT sum(length(toValidUTF8(materialize(repeat('é', 64)))))
FROM numbers(100000)
FORMAT Null
</query>
<query>
SELECT sum(length(toValidUTF8(materialize(concat(repeat('é', 64), 'a')))))
FROM numbers(100000)
FORMAT Null
</query>
<!-- Long valid mixed UTF-8 input. The expression is materialized to avoid measuring row construction. -->
<query>
SELECT sum(length(toValidUTF8(materialize(repeat('aé中文🙂', 80)))))
FROM numbers(100000)
FORMAT Null
</query>
<!-- Invalid input should repair from the first error without rescanning the prefix. -->
<query>
SELECT sum(length(toValidUTF8(materialize(concat('\xFF', repeat('a', 127)))))) AS s
FROM numbers(100000)
FORMAT Null
</query>
<query>
SELECT sum(length(toValidUTF8(materialize(concat(repeat('a', 63), '\xE2\x82X', repeat('b', 64)))))) AS s
FROM numbers(100000)
FORMAT Null
</query>
<query>
SELECT sum(length(toValidUTF8(materialize(concat(repeat('a', 127), '\xE2'))))) AS s
FROM numbers(100000)
FORMAT Null
</query>
</test>