Convert a column's values to a clean Number, Text, or Date form, with a fallback for values that won't convert.
Description
Coerces a column's cell values into a clean form of the type you pick — Text, Number, or Date — rather than flipping an abstract schema flag. Converting to Number tolerates thousands separators like commas; converting to Date accepts most common real-world date formats. Values that can't convert either block the run with an error naming which cells failed, or get replaced with a fallback value you choose, depending on how you configure it.
Best For
Turning a numeric-looking text column (e.g. "1,234.50") into real numbers before an Aggregate or Sort step
Normalizing inconsistent date formats after a PDF extraction
Confirming a column is fully clean by letting invalid values raise an error instead of silently passing through
How to Use
1Connect your table
Pass any table into Change Type — every column gets its own row, pre-selected to whatever type its current values look like.
2Set the target type per column
Choose Text, Number, or Date for each column you want to convert.
3Decide how to handle failures
Leave the fallback off to have any unconvertible cell block the run with a detailed error, or turn it on and supply a fallback value to use instead.
Tips & Notes
Dates always normalize to YYYY-MM-DD
Whatever format the source date is in — MM/DD/YYYY, "Jan 5, 2024", etc. — converting to Date drops any time component and outputs a plain YYYY-MM-DD string.
Off by default means strict
With the fallback disabled, a single bad cell in a large column stops the whole run. Turn on the fallback if you'd rather patch bad values and keep going.
