PSeitz
PSeitz
as reported by a user on Discord, the output of the Debug on error is a useless: 'Any { .. }' switch to `Display` via to_string
# Problem Outline Currently, tantivy stores f64 values uncompressed 8 bytes per value in its columnar storage. Generally float numbers are unsuitable for bitpacking, which requires a different approach, than...
Currently the tokenizer api generates a lot of allocations. For every Text encountered `text_analyzer::token_stream()` is called ```rust impl TextAnalyzer { /// Creates a token stream for a given `str`. pub...
- [X] Aggregations should release memory in `AggregationLimits` when they are finished - [ ] Term Aggregations could prune regularly results while collecting when reaching a threshold
Currently overlapping ranges are unsupported, but they have a valid use case, e.g. a custom facet for logs * Last Hour * Last 24 Hours * Last Week
JSON fields include type detection for numerical types on columnar (not sure about the term dictionary). It would be nice to add date detection to it. This should probably have...
``` SLOW [>120.000s] tantivy::failpoints test_fail_on_flush_segment_but_one_worker_remains SLOW [>180.000s] tantivy::failpoints test_fail_on_flush_segment_but_one_worker_remains SLOW [>240.000s] tantivy::failpoints test_fail_on_flush_segment_but_one_worker_remains SLOW [>300.000s] tantivy::failpoints test_fail_on_flush_segment_but_one_worker_remains SLOW [>360.000s] tantivy::failpoints test_fail_on_flush_segment_but_one_worker_remains SLOW [>420.000s] tantivy::failpoints test_fail_on_flush_segment_but_one_worker_remains SLOW [>480.000s] tantivy::failpoints test_fail_on_flush_segment_but_one_worker_remains SLOW...
The list provides an overview on which aggregations are missing in tantivy to reach feature parity with Elasticsearch. ## Bucket Aggregations ### Supported Terms Only text,u64, i64, f64 supported currently....