Open bobbyhubbard opened 9 years ago
+1
I am having the same issue with database errors that contain objects using quoted identifiers. My input data looks like "some exception querying ""sometable"" occurred" which as was mentioned is valid csv according to RFC4180 (paragraph 7).
+1
While attempting to parse a perfectly legit CSV, we're getting exceptions like the following:
As you can see, our csv field has embedded html which requires the use of the double quote in its content... necessitating two double quotes in order to differentiate between the field start/end and content inside the field. The RFC explains it better. Per RFC4180 (paragraph 7),
The logstash csv filter doesn't seem to account for the double double-quote scenario and therefore ends a field prematurely and spits out these malformed csv errors. I honestly haven't had a time to dig into the logstash code but I'm thinking this is a bug.
This is forcing us to use the elasticsearch-csv-river when we'd rather move to logstash. Has this been identified before? Any suggested workarounds?