Snowflake 的 REGEXP_SUBSTR 与 Python 连接器

Question

I am building a query with regex using the web UI, something like this:

SELECT uid, REGEXP_SUBSTR(PAGEPAGEPATHLEVEL3, '/(\\d+).*', 1, 1, 'e') as listing_id, SUM(TOTALHITS) as hits
FROM ga
WHERE PAGEPAGEPATHLEVEL2 = '/sales/'
GROUP BY (uid, listing_id)

And it works perfectly. However, once I pass the same query via python/pandas connection

pd.read_sql_query(query, con=con)

it also works, but all cells in listing_id are empty, I assume something In the regex should be escaped, but can't find any documentation on that

Answer 1

You need to escape the backslashes in your Python code. Add a r before your string:

query = r"""
SELECT uid, REGEXP_SUBSTR(PAGEPAGEPATHLEVEL3, '/(\\d+).*', 1, 1, 'e') as listing_id, SUM(TOTALHITS) as hits
FROM ga
WHERE PAGEPAGEPATHLEVEL2 = '/sales/'
GROUP BY (uid, listing_id)
"""

or double the backslashes:

query = """
SELECT uid, REGEXP_SUBSTR(PAGEPAGEPATHLEVEL3, '/(\\\\d+).*', 1, 1, 'e') as listing_id, SUM(TOTALHITS) as hits
FROM ga
WHERE PAGEPAGEPATHLEVEL2 = '/sales/'
GROUP BY (uid, listing_id)
"""