Yiyuan Liu

rt: optimize spawn functions

2

## Motivation Tokio doesn't peform well when spawing large futures. There are many function calls on spawn code path. Futures are passed as parameter values between these functions. Compilers can't...

Process termination.

1

Processes may terminate in a number of ways, eg SIGTERM, SIGKILL, machine down. Currently, Madsim does not simulate processes termination very well. Madsim provides `kill` function to terminate a host,...

RequestContext changed after co_await.

2

According to documentation, task coroutine is RequestContext-aware and will capture the current RequestContext at the time the coroutine function is either awaited or explicitly started and will save/restore the current...

Bug in performance test.

1

During performance test, I got some strange results. I noticed that in `class PerformanceTest`, it create a tmp buffer and copy data into tmp buffer before run sort. But sort...

Significant performance degradation when sorting strings with common prefix

1

### What happens? Env: - DuckDB main branch - 96 x 2 CPUs - 2TB memory - DuckDB is configured to use 64 cores and 1TB memory. ``` SET threads...

Improve performance of memory usage counters

3

In some cases, concurrent updates to the global memory usage counter can cause performance issues(#12749). This PR optimizes the performance of the memory usage counter but sacrifices some accuracy. A...

Add support for specifying columns in read_parquet.

2

Currently, DuckDB's projection pushdown optimization cannot work within structs. When accessing parquet files that contain structs, the entire struct will be read even if users only want some fields within...

Incorrect query result when using random value in where clause.

2

### What happens? The result of this query is incorrect, 0.5473918418865651 < 0.9. ``` select random() as num from range(20) where num > 0.9; ┌────────────────────┐ │ num │ │ double...