Support special characters for similarity queries #1182
Conversation
|
@procrastinatio please review |
|
Hmm, do not understand why you have to convert to ascii, the column type in the model should be |
| match.group(2).replace('\'', '\\\''), | ||
| match.group(3) | ||
| )) | ||
| where = where.encode('ascii', 'replace') |
There was a problem hiding this comment.
Why do you need to convert to ascii ?
There was a problem hiding this comment.
Because sqlAlchemy uses ascii to build plain SQL requests (this is where it fails).
replace argument replaces all the non-ascii characters with ?
There was a problem hiding this comment.
Not really up to date here, but e.g. ü is a non-ascii character, but doesn't cause any problems?
|
The error is unrelated to encoding errors, it is simply a matter of properly escape special characters in the partial SQL request. Normally, it is best to leave SQLAlchemy do the job. |
|
OK. |
|
and |
|
I know why prod and dev are behaving diffrently. pyramid_debugtoolbar is bugged see Pylons/pyramid_debugtoolbar#60 |
This PR is for: geoadmin/mf-geoadmin3#2009
At some point the where
unicodewill be converted to bytestrand will fail because for some reason we have non-ASCII unicode strings.The famous
UnicodeError...Here we
\and telling postrgres to escape it with E)strand replace unmapped characters with?that we finally replace with%It's a workaround but it works fine.
If anybody has a better solution/understanding of what's going on please feel free.
Current Link in prod
Deployed Branch in test