Ariel Shtul
Ariel Shtul
`idx` is `uint_t` and therefore always positive
I have several thread pools that might be created and destroyed and different times. Is there any reason not to include `threads_keepalive` in `thpool_` struct?
This has been added to master in the contains/wildcard PRs. #2774 #2886 RED-81612
In case many indexes were created with unique prefixes and dropped, the `SchemaPrefixNode` was not removed from the trie.
Currently, if a document is missing the sortby field, after the heap compare function assigned to it `-1` , it might change the sign to `1` if the sortby order...
and wrap with a macro that checks it is not `RSDummyCtx` on the coordinator.
On benchmark for GEO, an edge case was presented, where the iterator reached `EOF` with no result loaded into `res` due to internal filtering of distance, and therefore `res` was...
For example, ``` FT.CREATE idx SCHEMA t TEXT HSET doc1 t 1 HSET doc2 somethingelse 2 FT.SEARCH idx * SORTBY t ``` Result: `[2, 'doc2', ['somethingelse', '2'], 'doc1', ['t', '1']]`
This is an experiment to support glob matching. In addition to the functionality of wildcard (?, *), it supports list of chars ([ABC], [0-9], [^XYZ]). Initial performance tests show significant...
# consider sample index & documents ``` FT.CREATE userDetailsIdx ON JSON SCHEMA $.user.name AS name TEXT $.user.place AS place TEXT $.user.skills AS skills TAG $.user.age as age NUMERIC JSON.SET userDetail1...