I found an issue when the Bulk update SQL use CAST as CHAR, which doesn't work in postgresql. We need to CAST as VARCHAR. Then it works. The patch is only on the taxonomy. I am not certain if this works on mysql. So it is merely and example of a fix in one of the SQL statement. Perhaps someone know the two db's a bit better and can offer a complete solution.
| Comment | File | Size | Author |
|---|---|---|---|
| pathauto_taxonomy_pgsql.patch | 898 bytes | pearcec |
Comments
Comment #1
gregglesThanks for the patch!
Which version of PostgreSQL are you using?
We did lots of work on this with pgsql folks in #212327: cast the concatenated join value to be able to use the url_alias index so I'm hesitant to change this.
Comment #2
greggles...
Comment #3
pearcec commentedUnfortunately at the moment our dev environment is 7.4. But production is 8.2. I am almost certain I ran into this issue on 7.4.
I don't know when or if I will run into this issue again. But should we change the status to "patch (code needs review)" and just let it sit in case someone else has an issue? I don't feel compelled to make 7.4 work.
Comment #4
gregglesIf this only impacts 7.4 (I imagine that is the case) then I think we might want to just dismiss it...
But, for now, yes, let's leave the issue open.
Comment #5
justin3 commentedWorks for PG 8.2 !
I had the issue with bulk update returning just /taxonomy/term for a new alias, then stopping- ie not updating any more taxonomy aliases even on future updates.
Then I switched to pathauto 6.x-1.x-dev - which fixed that error, but pathauto bulk update was still not adding new taxo terms- until this patch fixed it.
thanks!
Comment #6
justin3 commentedJust had to make this same fix for pathauto_node.inc to get the bulk updates to continue for nodes
line 100 - change CHAR to VARCHAR
Comment #7
gregglesAnyone use pgsql8.3 to be able to test it? I feel confident that this works for some people...
Comment #8
dave reidDuplicate of #319070: Provide a wrapper for cross-db and cross-field-type concatenation (PosgreSQL)
Comment #9
dave reid