har.fyi has documented two of the tables already but given BigQuery appears to support column descriptions, really we should co-locate the descriptions in the place we're actually querying them.
It's difficult as an irregular user of the HTTPArchive to reverse-engineer what the columns include or even which table I should be querying. For the latter, I'm not sure if BigQuery supports table descriptions/comments but that would naturally be useful too.
It may be that this schema could then become the source of truth for har.fyi to pull from rather than maintaining two copies.
har.fyi has documented two of the tables already but given BigQuery appears to support column descriptions, really we should co-locate the descriptions in the place we're actually querying them.
It's difficult as an irregular user of the HTTPArchive to reverse-engineer what the columns include or even which table I should be querying. For the latter, I'm not sure if BigQuery supports table descriptions/comments but that would naturally be useful too.
It may be that this schema could then become the source of truth for har.fyi to pull from rather than maintaining two copies.