The Integrity Text Search / Context Search functionality is integral functionality in Integrity.
In the Integrity Client there exist different text search functions, which all work a bit differently though:
- Text filter in document view
- Attribute filter on text attributes in document view
- Text filter in items view
- Attribute filter on text attributes in document view
- Search in document via CTRL-F
- Global text search in available in items view
- Query filters on text attributes
- Text search in a text attribute
Every text search function works a bit different and none of the work as expected:
Examples:
- Different results depending on content is in plain or rich text field - Sometimes you need LIKE and other database keywords, sometimes not
- Sometimes wild cards are allowed, sometimes not
- Sometimes boolean operators are allowed, sometimes not - Some characters (e.g. Underscore) need to be escaped ("\"), others not
- Some characters are ignored, e.g. ! , # , [], and @ (mostly undocumented)
- Format change in a search term is never found
- Sometimes the stop list comes into play etc. etc.
This said, text search does not work reliably, meaning in most cases the user does not get the results expected and does not even notice.
This is a direct violation of the ISO26262 and not acceptable in safety relevant development.
Expectations:
- All text searches need to work in the same way, of course differentiating between “Contains” and “Containing” text searches
- Optimize text search functionality documentation (Configuration and User Guide). Currently it seems like an incomplete puzzle. Configuration, UserGuide, Document Cases (KBs),..:
- Integrity properties configurations
- Database specific configurations (index/lexer, ignore/noise words, etc.) and how configurations affect text search. Database dependent (Oracle, MS SQL Server). Some Case Documents exists, but nobody knows or difficult to understand the whole story.
- Simple Text search to advanced ones searching for words containing special characters like underscore etc., random results up to 100% match.
- Explain the usage of database specific keyword i.e. LIKE, IS, =, etc.
When you can and when you have to use them. Depending on search result accuracy expected and special character usage in search strings
- Many users do not know the difference between “Contains” and “Containing” text searches. Explaining a bit more in the Integrity User Guide would help as well.
- Also document what is not possible regarding special characters, escaping or database keywords. This would save time and discussions as to text search result accuracy.
