babelfish-for-postgresql / babelfish_extensions / 12693842151

77%

push

github

Fix hashbytes behaviour for nvarchar input (#3259) (#3385)

PROBLEM: while having nvarchar input in hashbytes we were considering the encoding as input encoding in babelfish where as in tsql we use UTF16 encoding for nvarchar irrespective of input encoding.

RCA: we were considering varchar and nvarchar as same, whereas we should use input encoding for varchar and UTF16 encoding for nvarchar.

FIX: So we need to identify that if the input is nvarchar then we will do the UTF16 encoding.
In the function hashbytes, we were converting varchar/nvarchar to varbinary, so if we get input as nvarchar then we will first convert the input string to UTF16 encoding then convert to varbinary.
There we created the overloaded function and also select the best candidate among depending on the second argument(i.e varchar/nvarchar/varbinary).

Task: BABEL-4891

Signed-off-by: pranav jain <pranavjc@amazon.com>

39 of 39 new or added lines in 3 files covered. (100.0%)

46860 of 62524 relevant lines covered (74.95%)

167631.74 hits per line