Show HN: SPyQL – SQL with Python in the middle - Klaput News
Show HN: SPyQL – SQL with Python in the middle
3 by dmoura | 0 comments on Hacker News.
SPyQL ( https://ift.tt/3rJfzPE ) is SQL with Python in the middle, an open-source project fully written in Python for making command-line data processing more intuitive, readable and powerful. Try mixing in the same pot: a SQL SELECT for providing the structure, Python expressions for defining transformations and conditions, the essence of awk as a data-processing language, and the JSON handling capabilities of jq. How does a SPyQL query looks like? $ spyql “ IMPORT pendulum AS p SELECT (p.now() - p.from_timestamp(purchase_ts)).in_days() AS days_ago, sum_agg(price * quantity) AS total FROM csv WHERE department.upper() == 'IT' and purchase_ts is not Null GROUP BY 1 ORDER BY 1 TO json” < my_purchases.csv In a single statement we are 1) reading a CSV (of purchases) with automatic header detection, dialect detection, type inference and casting, 2) filtering out records that do not belong to the IT department or do not have a purchase timestamp 3) summing the total purchases and grouping by how many days ago they happened, 4) sorting from the most to the least recent day and 5) writing the result in JSON format. All this without loading the full dataset into memory. The Readme is loaded with recipes and there is also a demo video: https://vimeo.com/danielcmoura/spyqldemo Any feedback is welcomed! Thank you.
RELATED POST :
3 by dmoura | 0 comments on Hacker News.
SPyQL ( https://ift.tt/3rJfzPE ) is SQL with Python in the middle, an open-source project fully written in Python for making command-line data processing more intuitive, readable and powerful. Try mixing in the same pot: a SQL SELECT for providing the structure, Python expressions for defining transformations and conditions, the essence of awk as a data-processing language, and the JSON handling capabilities of jq. How does a SPyQL query looks like? $ spyql “ IMPORT pendulum AS p SELECT (p.now() - p.from_timestamp(purchase_ts)).in_days() AS days_ago, sum_agg(price * quantity) AS total FROM csv WHERE department.upper() == 'IT' and purchase_ts is not Null GROUP BY 1 ORDER BY 1 TO json” < my_purchases.csv In a single statement we are 1) reading a CSV (of purchases) with automatic header detection, dialect detection, type inference and casting, 2) filtering out records that do not belong to the IT department or do not have a purchase timestamp 3) summing the total purchases and grouping by how many days ago they happened, 4) sorting from the most to the least recent day and 5) writing the result in JSON format. All this without loading the full dataset into memory. The Readme is loaded with recipes and there is also a demo video: https://vimeo.com/danielcmoura/spyqldemo Any feedback is welcomed! Thank you.
RELATED POST :
Memuat...
Comments
Post a Comment