uber/queryparser

Column usage doesn't include hive struct accessors

Open

#11 创建于 2017年8月29日

在 GitHub 查看
 (1 评论) (0 反应) (0 负责人)Haskell (1,079 star) (145 fork)batch import
good first issue

描述

For tables where the majority of 'columns' are abstracted under a top-level column, e.g. body or msg, we don't surface the struct accessor as the column name, rendering analysis of these queries unusable.

Ideally, if we had a msg.name, msg.attrs.foo, msg.attrs.bar struct, we'd see the leaf struct names as the column name, i.e. msg.name, msg.attrs.foo, etc. Instead, we currently see msg for all of them.

贡献者指南

技术栈
sql
领域
databaseanalytics
议题类型
feature
难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。
3
预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。
1-2 days
活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。
stale
清晰度议题是否清楚说明期望改动、验收标准和下一步。
clear
前置要求
Basic understanding of SQL parsingFamiliarity with Hive struct syntax
新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。
20
研究方向
Investigate the column name extraction logic in the parser. The issue requests that when accessing struct fields like msg.name, the column name should reflect the leaf struct name. Look at how the parser currently retrieves column names and identify where the struct accessor is lost. Consider modifications to propagate the full path. See the issue at https://github.com/uber/queryparser/issues/11.