Data type detection in Tableau


When you connect a data source, Tableau automatically infers the type of each column of your data. Whether it's Number (decimal), Number (whole), String, Boolean, etc. A few questions about this:

1) Has Tableau ever misclassified one (or more) columns of your data with their automatic labeling? If so, would you please give details?

2) Do you feel that these data type options can be improved? For example, Tableau doesn't seem to be distinguishing between unordered categorical data and ordinal data.


Rohan Kadakia

Posted 2018-06-10T18:11:01.253

Reputation: 135

Question was closed 2018-06-13T16:11:32.527

This sounds like an experience/opinion poll. Those kinds of questions usually aren't a good fit here; see our [help/dont-ask]. Do you have an objectively answerable question? An example of an objectively answerable question might be "Is it possible for Tableau to misclassify a column?" or "What algorithm could we use to distinguish between unordered categorical data vs ordinal data?". Can you edit your question to make it suitable here? – D.W. – 2018-06-10T20:21:01.127

The objectively answerable question is question (1). I'm particularly looking for situations where Tableau consistently misclassifies data type. – Rohan Kadakia – 2018-06-11T00:30:48.303



  1. Yes, it sometimes does that. I think it depends on the data source. For RDBMS it is likely that it takes the data type from the database (so the column type is typically accurate). For text/csv, it probably does a scan of the top certain lines to guestimate the type. It is similar to the text import wizard of the SQL, etc. other similar tool. (A Tableau user of 4+ years)
  2. I am not sure if it is relevant to many of the Tableau users; many databases don't have that distinction either; I think it is more helpful for someone working in machine learning, etc.?

The Lyrist

Posted 2018-06-10T18:11:01.253

Reputation: 454