censhare Quick Search - Result Is Not as Expected (Caused by Using Stop Words)

Short description of the following content

This document will show you how to find out if a quick search result is caused by using stop words during indexing of embedded database values.

Important hints

Please find basic information about the quick search here

• In censhare Server Versions < 4.5 the usage of stop words is activated by default

• In censhare Server Versions > 4.5 the usage of stop words is deactivated by default

• After changes to the embedded database, a censhare Server restart, as well as a CDB rebuild is necessary

Question/Symptom

A quick search for "honorarfrei für yacht: buyout" results in 5,872 hits.
A quick search for "honorarfrei yacht: buyout" results in 70,000 hits, which is closer to the expected result.
There are 71,648 assets with the feature value "honorarfrei für yacht: Buyout" within the field test:pictstore.honorar.
This number of assets would be expected as a result of a quick search.
What is the cause?

Answer/Solution

This is an effect of using the stop word list. The word "für" is a default (german) stop word and will be skipped while indexing, but it can't be ignored during the query, because it could be the beginning of a hit word (for example the german word "Fürsorge" (en: care)).
The decisions of the query engine can be traced by using the server protocol with log level "FINER" or "FINEST".
If the stop word list shouldn't be used at all it can be deactivated with the attribute stopword-list="false" within the index configuration at censhare Admin Client under Configuration | Service | Embedded database.

Note: For censhare Server versions < 4.5 this must be done directly within the xml configuration file.