About Files for Datasets

You can upload Microsoft Excel spreadsheets (XLSX and XLS), CSV files, and TXT files and then create datasets from them. The maximum file size you can upload is 250 MB and the data column limit for a single file is 250 columns.

You must structure an Excel spreadsheet in a data-oriented way with no pivoted data. These rules apply for Excel tables:

• Tables must start in Row 1 and Column 1 of the Excel file.
• Tables must have a regular layout with no gaps or inline headings. An example of an inline heading is one that's repeated on every page of a printed report.
• Row 1 must contain the names of the columns in the table. For example, Customer Given Name, Customer Surname, Year, Product Name, Amount Purchased, and so on. In this example:
  • Column 1 has customer given names.
  • Column 2 has customer surnames.
  • Column 3 has year values.
  • Column 4 has product names.
  • Column 5 has the amount each customer purchased for the named product.
• The names in Row 1 must be unique. If two columns hold year values, then you must add a second word to one or both of the column names to make them unique. For example, if you have two columns named Year Lease, then you can rename the columns to Year Lease Starts and Year Lease Expires.
• Row 2 and greater must contain the data for the table, and those rows can't contain column names.
• Data in a column must be of the same type because it's often processed together. For example, the Amount Purchased column must contain only numbers (and possibly nulls), enabling it to be summed or averaged. The Given Name and Surname columns must contain text values because they might be concatenated, and you might need to split dates into months, quarters, or years.
• Data must be at the same granularity. A table can't contain both aggregations and details for those aggregations. For example, suppose that you have a sales table at the granularity of Customer, Product, and Year that contains the sum of Amount Purchased for each Product by each Customer by Year. In this case, you wouldn’t include invoice level details or daily summary values in the same table, because the sum of the Amount Purchased values wouldn't be calculated correctly. If you must analyze at the invoice level, the day level, and the month level, then you can do either of the following:
  • Have a table of invoice details: Invoice Number, Invoice Date, Customer, Product, and Amount Purchased. You can roll these up to the day, month, or quarter.
  • Have multiple tables, one at each granular level (invoice, day, month, quarter, and year).