babelfish-for-postgresql / babelfish_extensions / 10109171598

76%

BABEL_5_X_DEV: 77%

Build Type

push

github

Committed by web-flow

Commit Message

Support strpos & replace for non deterministic collation (#2716)

Support strpos, charindex and replace for Accent insensitive collations in Babelfish. We use ICU functions such as
usearch and ucol_strcoll to implement this support.

sys.strpos()/ sys.charindex(): we use usearch to find the first available match in source string. If collation is CS_AS then
we also double check the match using ucol_strcoll. This is needed since usearch can return false positive for case
sensitive search when collation is colstrength primary. If ucol_strcoll fails for the match, we skip to the next match.

sys.replace: Similar to strpos we find all the matches using usearch, double check it using ucol_strcoll for CS_AS. For
the matched portion of the string we copy the replacement string instead while copying the remaining portion from the
original source string.

Engine PR: babelfish-for-postgresql/postgresql_modified_for_babelfish#402
Task: BABEL-4850
Signed-off-by: Tanzeel Khan <tzlkhan@amazon.com>

Run Details

136 of 146 new or added lines in 3 files covered. (93.15%)

1 existing line in 1 file now uncovered.

43713 of 59398 relevant lines covered (73.59%)

113402.06 hits per line