Indexing and Searching - partial string search support?

Are there any plans to enable partial word searches in the Apache Lucene search based Index Query node (knime labs extensions, indexing & searching)? The node details for wildcard search indicate wildcards are not allowed at beginning of the query, i.e. *test not allowed, and the results indicate whole-word only matches, i.e. “test” matches only “blah test blah” and doesn’t match “blah retest blah” or “blah untested blah”. The regex doesn’t seem right to me either, in that /test/ should match partial words but does not.

Per this StackOverflow thread, partial word searches are possible:

RE: RegEx, I was expecting this regex pattern matching but didn’t observe it on testing in Knime 3.6.0
https://regexr.com/425bj
(only whole words returned)

Indexing & Searching info here: https://www.knime.com/indexing-and-searching
Node description: https://nodepit.com/node/org.knime.ext.lucene.nodes.lucenequery.LuceneQueryNodeFactory

The context in this case is I’m trying to search an entire table space for any partial matches to specific terms for further review as this is an unknown source and unsure which of hundreds of columns to check for such terms. Trying to use index&search as an alternative to looping through each column and collecting/combining RowIDs into a row filter node.

Thanks.

Hi,
if you want to find words that contain test you need to use the following regular expression /.test./ instead of /test/ see attached workflow.
Bye
Tobias
IndexAndSearchRegex.knwf (8.1 KB)

1 Like