Data cleaning functions — extract, classify, tag, format

A family of functions in GPT for Sheets for turning messy text into structured columns. All of them take a cell as input and are built to be dragged down thousands of rows.

The functions

Function	What it does
=GPT_EXTRACT(text, to_extract)	Pull entities out of text: emails, names, prices, SKUs.
=GPT_CLASSIFY(value, categories)	Assign one category from your list.
=GPT_TAG(value, tags)	Apply multiple tags from your list.
=GPT_FORMAT(input, target_format, [source_format])	Normalize dates, phones, addresses, casing.
=GPT_SPLIT(text, split_by) / =GPT_HSPLIT()	Split text into rows / columns by meaning.
=GPT_APPLY(text, [task])	Apply an edit like “fix grammar and spelling”.
=GPT_SUMMARIZE(text, [format])	Summarize into a given shape (“three sentences”).

Examples

Extract the email from a messy signature in A2:

=GPT_EXTRACT(A2, "email address")

Classify leads by industry with your own taxonomy in $F$1:

=GPT_CLASSIFY(A2, $F$1)

Normalize phone numbers:

=GPT_FORMAT(A2, "+1 (XXX) XXX-XXXX")

Try it

Copy the template or install GPT for Sheets. Recipe: data cleaning template.