1 of 1 people found this helpful
Do you have a schema.ini in the same folder as the files? If so, delete it and try reconnecting. It is automatically created, but if it's been modified it could lead to the issues you are seeing.
I tried that, as that seems to have caused me problems in the past, but it seems to persist in this case. Poking through the data, it looks like the first several dozen (maybe over 100) rows all have no data and are showing as "0". How many records does Tableau consider before assigning a field's data type?
"How many records does Tableau consider before assigning a field's data type"
It's not T fault, is the MS Jet Engine the main cause of issue. the # of records are default ones used by the Jet Engine. I think that you could instruct it to read more records, using a special value placed in .ini file.
MaxScanRows The number of rows to be scanned when guessing the column types. If set to 0, the entire file will be searched. The default is 25.
I found the answer here
If you deal with .txt .csv or .tsv files try to import the file in MS Access, run some checks and then use Tableau to import from Access. It's a more robust way of having data in Tableau Data Engine.
Ding, ding, ding, we have a winner. Other alternatives are using a schema.ini to force float, or adding a dummy line of data with a 0.0.
Per your MaxScan comment, I've never been able to depend on that - JET has still screwed up.
And now I must go back to my IT group - for some reason Access has blocked me from importing txt files for some unknown reason. I think if they can get me back to the point where I can use that as a starting point I'll have this problem solved. Thanks all for the input.
Truth be told, I'm not sure I want to search the entire record since there are about 65 million rows...
With that many rows I would use a schema to force float then.