Closed van-fs closed 4 years ago
Hey @van-fs thanks for pointing this out. We actually don't officially support CSV and local storage yet because we'd like to have consistent headers (or at least column ordering) between files. I'll be working on a change to officially support this soon.
The reason that the header is not included is because when loading data into the database (both BigQuery and Redshift), we aren't specifying that a header should be ignored. So in order to support writing headers (which we should do), we need to update the database loading to ignore the header as well.
When using local storage:
The written CSV files lack headers (i.e. column names). The quick fix appears to be simply adding a
Write
with the existingtableColumns
inProcessFilesByDay
andProcessFilesIndividually
.Is there any reason that we should not write headers in this scenario? I haven't submitted a PR as I think we might make this configurable if headers should not be present in CSVs to work correctly with other adapters.