Re: [Tiki-devel] Issue with innoDB MySQL search results for "he" (but not for "fr" or "en")
Bsfez Tiki via TikiWiki-devel <[email protected]>
| Newsgroups | gmane.comp.cms.tiki.devel |
|---|---|
| Message-ID | <[email protected]> |
Hi Jonny,
Length is good:
MariaDB [information_schema]> SHOW VARIABLES LIKE 'ft_min_word_len';
+-----------------+-------+
| Variable_name | Value |
+-----------------+-------+
| ft_min_word_len | 2 |
+-----------------+-------+
1 row in set (0.002 sec)
But I just find the culprit !!!
"sort mode"
—
{list max="8"}
{filter type="file"}
{filter field="gallery_id" content="18"}
{sort mode="randommode" modes="title_asc, object_id, creation_date_nasc"}
{filter field="title" content="Kindergarten AND Screen AND {/literal}{$prefs.language}{literal}"}
—
The same code will work fine with fr or en. (language pref or hardcoding it : ...Screen AND en…)
If I do
{sort mode="randommode" modes="title_asc, object_id, creation_date_nasc"}
{filter field="title" content="Kindergarten AND Screen AND he"}
Result will show ONLY "fr" filename or ONLY "en" filename but never "he" filename.
Only when I remove the sort block "he" filename are find correctly.
huuuu 👀
I have not enough knowledge to understand how this could be possible...
Bernard
> On 21 Jul 2021, at 11:41 , Jonny Bradley via TikiWiki-devel <[email protected]> wrote:
>
>
> Yes, he is generally a pronoun in English, and sadly (according to lib/core/Search/ContentSource/FileSource.php:59) all files are indexed as language=unknown... can't think of an easy way round that - maybe have different galleries for the various langs?
>
> Meanwhile, what is your ft_min_word_len there? I think it defaults to 4, so maybe you need to set that to 2 only? https://dba.stackexchange.com/a/40823
>
> jonny
>
>
>
>
>> On 21 Jul 2021, at 09:18, Bsfez Tiki via TikiWiki-devel <[email protected]> wrote:
>>
>> Hello,
>>
>> May be someone can help with an issue I have since a few versions for on Tiki22. (was working previously)
>>
>> In a file gallery I have files with in their filenames "en", "fr" or "he".
>>
>> If I perform a search and filter files with "en" and "fr" in the filename I got results.
>> If I perform a search and filter files with "he" no results found.
>>
>> I tried to locate and check the stopwords list used as "he" could be seen as a "pronoun".
>>
>> In the Tiki admin the stopwords (never touched) field doesn’t show a "he" (not sure why there is a "s" and a "t"
>>
>> And it is the same in the innoD stopword list (no "he")
>>
>> MariaDB [information_schema]> SELECT * FROM INFORMATION_SCHEMA.INNODB_FT_DEFAULT_STOPWORD;
>> +-------+
>> | value |
>> +-------+
>> | a |
>> | about |
>> | an |
>> | are |
>> | as |
>> | at |
>> | be |
>> | by |
>> | com |
>> | de |
>> | en |
>> | for |
>> | from |
>> | how |
>> | i |
>> | in |
>> | is |
>> | it |
>> | la |
>> | of |
>> | on |
>> | or |
>> | that |
>> | the |
>> | this |
>> | to |
>> | was |
>> | what |
>> | when |
>> | where |
>> | who |
>> | will |
>> | with |
>> | und |
>> | the |
>> | www |
>> +———+
>>
>> Does anyone can help me finding why "he" is excluded from search ?
>>
>>
>> Thanks,
>> Bernard
>>
>>
>>
>>
>> <Screen Shot 2021-07-21 at 11.06.44 .png>
>>
>> <Screen Shot 2021-07-21 at 10.52.28 .png><Screen Shot 2021-07-21 at 10.52.53 .png><Screen Shot 2021-07-21 at 10.53.02 .png><Screen Shot 2021-07-21 at 11.08.02 .png>
>>
>> _______________________________________________
>> TikiWiki-devel mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/tikiwiki-devel
>
>
>
> _______________________________________________
> TikiWiki-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/tikiwiki-devel
_______________________________________________
TikiWiki-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/tikiwiki-devel