Alexander Kuzmenkov
Alexander Kuzmenkov
This PR adds vectorized computation of text equality for deterministic collations, and case-sensitive LIKE for UTF-8 database encoding. Benchmarking shows up to 3x speedups: https://grafana.ops.savannah-dev.timescale.com/d/fasYic_4z/compare-akuzm?orgId=1&var-branch=All&var-run1=3171&var-run2=3178&var-threshold=0&var-use_historical_thresholds=true&var-threshold_expression=2.5%20%2A%20percentile_cont%280.90%29&var-exact_suite_version=false Prerequisites: - [x] https://github.com/timescale/timescaledb/pull/6272 -...
The compressed qual pushdown will look at presence of metadata columns in the compressed chunk, not at whether the uncompressed column is an orderby. This way, the decision on which...
It is about 10x faster than row-by-row decompression, so if we're going to decompress 10x less rows than the batch size, it's more efficient to decompress row-by-row. In tsbench, this...
This simplifies passing the columnar data out of the DecompressChunk to Vectorized Aggregation node which we plan to implement. Also this should improve memory locality and bring us closer to...
Apply the same approach we use for chunk exclusion: try to cast the constant to the same type. Disable-check: force-changelog-file
Late materialization on the level of postgres tuple protocol (tts_getsomeattrs) doesn't do much for performance: https://grafana.ops.savannah-dev.timescale.com/d/fasYic_4z/compare-akuzm?orgId=1&var-branch=All&var-run1=3236&var-run2=3235&var-threshold=0&var-use_historical_thresholds=true&var-threshold_expression=2.5%20%2A%20percentile_cont%280.90%29&var-exact_suite_version=false The only exception is some text filter query which is going to be covered...
This can avoid materializing extra columns.
Add ANALYZE after compression of `metrics_compressed` in shared test setup. ``` ci_stats=# select distinct on (log_contents) count(*) over (partition by log_contents), * from test natural join log where test_name like...
We don't account for them at all currently, which leads to weird planning behavior, like the filters not influencing the position of relations in join. We don't normally have the...
Investigate problems described in this issue https://github.com/timescale/timescaledb/issues/6310#issuecomment-1811491927 Parts: - [x] https://github.com/timescale/timescaledb/pull/6422 - [x] https://github.com/timescale/timescaledb/pull/6419 - [x] https://github.com/timescale/timescaledb/pull/6323 - [x] https://github.com/timescale/timescaledb/pull/6330 Partial aggregation is broken in main on some very basic...