Hacktoberfest 2026：维护者为十月标记出来的 issue，仍然开放、适合新手。 浏览 Hacktoberfest issue

Perf: update open_era5.py to use read_xarray_table with filter pushdown and lazy loading

未关闭
  • 哪里不对？
#139 0 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看

维护者通常 1 天内回复

还没有人认领这个 Issue。

评估

难度
3/5
预计耗时
1-2 天
新手友好度
52/100
Issue 类型
功能
描述清晰度
描述清楚
活跃度
停滞
技术栈
google-cloud, python
领域
data-engineering, performance

调研方向

Start with perf_tests/open_era5.py and compare its existing read_xarray call with the proposed read_xarray_table flow. Add measurements for registration, a time filtered query, and a full scan, then run the benchmark with GCP authentication available and check that the output reports timings for each workload.

由索引模型根据 Issue 内容生成。

描述

Context

perf_tests/open_era5.py currently uses the older read_xarray API, which returns an eager pa.RecordBatchReader. It also only prints the schema, making it impossible to benchmark real query performance.

Current state

# open_era5.py
era5_ds = xr.open_zarr("gs://gcp-public-data-arco-era5/...", chunks={"time": 240, "level": 1})
era5_wind_df = xql.read_xarray(
    era5_ds[["u_component_of_wind", "v_component_of_wind"]]
)
print(era5_wind_df.schema)

Issues:

  1. Uses read_xarray instead of read_xarray_table — loses filter pushdown and lazy partition loading
  2. Only inspects schema, never runs a query — no way to measure actual query performance
  3. Does not demonstrate the partition pruning benefit for time-range queries

Proposed update

# open_era5.py
import time
from datafusion import SessionContext
import xarray as xr
import xarray_sql as xql

era5_ds = xr.open_zarr(
    "gs://gcp-public-data-arco-era5/ar/1959-2022-full_37-1h-0p25deg-chunk-1.zarr-v2",
    chunks={"time": 240, "level": 1},
)
wind_ds = era5_ds[["u_component_of_wind", "v_component_of_wind"]]

# read_xarray_table: lazy, supports filter pushdown
table = xql.read_xarray_table(wind_ds, chunks={"time": 240})

ctx = SessionContext()
ctx.register_table("era5_wind", table)

# Filter pushdown: should prune most partitions
t0 = time.perf_counter()
result = ctx.sql(
    "SELECT AVG(u_component_of_wind), AVG(v_component_of_wind) "
    "FROM era5_wind WHERE time > '2020-01-01'"
).to_arrow_table()
print(f"Filter query: {time.perf_counter() - t0:.2f}s  {result}")

Goals

  • Serve as a real-world benchmark for ERA5 (the primary large-scale use case)
  • Show filter pushdown pruning thousands of partitions
  • Compare registration time, filter query time, and full-scan time
  • Can be run with py-spy to generate flame graphs once GCP auth is available
主要语言
Python
星标
142
派生
19
平均合并
1 天 9 小时
30 天内合并 PR
9

环境准备

  • 没有 Dockerfile 或 Docker Compose 文件
  • 没有 Pull Request 模板
  • 阅读贡献指南

从这里开始

  1. 先读完整个 Issue，再读项目的贡献指南。
  2. 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
  3. Fork 仓库，在一个分支上完成修改。
  4. 提交 Pull Request，并在描述里引用这个 Issue 编号。

xqlsystems/xarray-sql 的其他 Issue

查看 xqlsystems/xarray-sql 的全部 Issue

相似的 Issue

更多 Python Issue

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。