AWS Glue data quality dynamic rules not working

I tried AWS Glue data quality dynamic rules in my AWS Glue pipeline. I wrote below rule RowCount > avg(last(3))

Then I processed 3 csv files with 1000,10000 and 100 rows. Then in 4th run I again processed file with 100 rows, expecting this rule to fail. But it didnt fail. I got below pass results :

RowCount > avg(last(3)) Rule passed Dataset.*.RowCount: 100.00

Gonzalo Herreros 专家
3 个月前
Yep, it sounds it thinks the average is 100, not sure why. What did it say on the previous evaluations?
How did you make it process different files on different runs?, do you have any count that the job actually read them all? (maybe it has thrown away invalid rows)
manish05485
3 个月前
This worked on 5th time. I think the issue is, in my first 3 runs this rule was not there. I added this rule in 4th run only. So it didnt took earlier runs into consideration. But this thing is not mentioned anywhere in documentation. There is no mention how Glue stores 'state' of runs.

主题

分析数据库

标签

AWS Glue 提取变换和加载数据

语言

English

manish05485

已提问 3 个月前171 查看次数

1 回答

最新
投票最多
评论最多

That's a fair point about clarifying the documentation, the history is based on the historical metric, if the rule wasn't applied then the metric wasn't created, the rule cannot backprocess the data to calculate the previous history if the rule wasn't applied.

专家

Gonzalo Herreros

已回答 2 个月前

manish05485
2 个月前
HI can you please see if you have any thoughts on this question ?https://repost.aws/questions/QU88DiiUZ1STmIjySismV2Ow/how-does-aws-glue-data-quality-custom-sql-work-with-no-unique-column

AWS Glue data quality dynamic rules not working

相关内容