Since search_api database backend is default as of #1140994: Solr search vs database search I got the following error on a recruiter site while indexing:

Indexing a maximum number of 50 items for index Job search.
WD search_api_db: An overlong word (more than 50 characters) was encountered  [warning]
while indexing, due to bad tokenizing. It is recommended to enable the
"Tokenizer" preprocessor for indexes using database servers. Otherwise, the
service class has to use its own, fixed tokenizing.
WD search_api_db: An overlong word (more than 50 characters) was encountered  [warning]
while indexing, due to bad tokenizing. It is recommended to enable the
"Tokenizer" preprocessor for indexes using database servers. Otherwise, the
service class has to use its own, fixed tokenizing.
WD search_api_db: An overlong word (more than 50 characters) was encountered  [warning]
while indexing, due to bad tokenizing. It is recommended to enable the
"Tokenizer" preprocessor for indexes using database servers. Otherwise, the
service class has to use its own, fixed tokenizing.
Drush command terminated abnormally due to an unrecoverable error.            [error]
Error: Unsupported operand types in
/.../search_api_db/service.inc,
line 506

enabling the solr backend works as a workaround to the the problem:
drush en recruiter_search_solr

Comments

mh86’s picture

Status: Active » Fixed

Partly fixed that problem by enabling the tokenizer for the database backend. The tokenizer will help to split words with whitespaces. Nevertheless, words longer than 50 characters can't be correctly indexed by the database backend (see #1117074: Indexing does not complete for more info). Fortunately words longer than 50 characeters are very uncommon for most languages.

Status: Fixed » Closed (fixed)

Automatically closed -- issue fixed for 2 weeks with no activity.

menatito’s picture

Title: Indexing fails with database backend » Indexing fails with database backend - search_api_db
Issue summary: View changes

An overlong word (more than 50 characters) was encountered while indexing, due to bad tokenizing. It is recommended to enable the "Tokenizer" preprocessor for indexes using database servers. Otherwise, the service class has to use its own, fixed tokenizing.

dboeger’s picture

These warnings are getting triggered in our site by links in the body field. Are there any suggestions about how to handle links in the body field so they don't trigger these errors?