Re: [BUGS] Notes about behaviour of SIMILAR TO operator

2003-11-21 Thread Tom Lane
Adam Buraczewski <[EMAIL PROTECTED]> writes: > ('a' similar to 'a' escape null) is true (should be unknown!) Yeah, you are right; this is because we are overloading a "null" second parameter to mean "the ESCAPE part wasn't present", which in hindsight wasn't such a hot idea. > I think that

Re: [BUGS] Notes about behaviour of SIMILAR TO operator

2003-11-21 Thread Adam Buraczewski
Hallo Tom, I decided to improve similar_escape() function during the weekend. Thank you very much for the excerpt from SQL standard (I think this is much more complete than the text I found in a Working Draft from August 1994). However, there are some more issues I'd like to make clear before fun

Re: [BUGS] Notes about behaviour of SIMILAR TO operator

2003-11-20 Thread Tom Lane
Adam Buraczewski <[EMAIL PROTECTED]> writes: > ... for example the pattern 'a|z' (which should match single 'a' or 'z' > characters only, according to SQL spec) is converted into POSIX > regular expression in the form of '^a|z$' which matches all strings > beginning with 'a' ('abcdef' for example)

[BUGS] Notes about behaviour of SIMILAR TO operator

2003-11-20 Thread Adam Buraczewski
Hallo pgsql-bugs, I was recently playing with SIMILAR TO operator (in both PostgreSQL 7.3.x and 7.4) and discovered that sometimes it behaves not the way I expected. Since this operator is quite new for me (I am familiar mainly with perl, grep, sed and emacs regexps and their syntax is slightly d