Repository navigation
readCSV does not produce an array convertable to query when firstRowHeader is off #443
Description
Activity
CFML queries must have column names, which is why the library will generate them if the source spreadsheet doesn't have them.
But CSVs don't have to have headers/columns and I think the result should represent the contents of the CSV as closely as possible.
I understand the problem if you ultimately want a query object, but in that case you should probably use csvToQuery() instead.
Your suggestion would be a breaking change. You might think it would be safe, but other folks might be relying on those empty columns, which as I say matches the csv input.
Having said that, I have wondered whether to include the query de/serialization as an option, which would mean
csvToQuery()could be deprecated... In other words addintoAQuery()as a chainable option. That would require auto generation of columns if they were missing. Thoughts?csvToQuery()does not allow me to skip rows.Jumping between meetings.
intoAQuery()seems like a workable solution.OK, John, I'll see what I can do.
Reacted by Daemach@Daemach I've added support for
intoAQuery()to the dev branch. Can you test it out please? See examples in the test suite:spreadsheet-cfml/test/specs/readCsv.cfm
Line 277 in f3634cd
It works. Thank you!
Reacted by Julian Halliwell
The normal functionality of spreadsheet-cfml is to return a query with
column1,column2,columnNas headers if the first row isn't used. But readCSV produces an empty column array in this case, which preventsdeserializeJson( serializeJson( result ), false )from working properly. An empty column array would be ignored normally, so it seems safe to populate this withcolumn1,column2,columnNas headers if the first row isn't used.