CSV column mapping
Every spreadsheet import works the same way. You match your columns to ours, check the rows, and only then are they added.
Last reviewed
Matching columns on every spreadsheet import
How an import works
Whichever list you bring in, the screen is the same, with four steps along the top: Upload, Match columns, Check and Done. Nothing is added to your practice until the last step. The import screens are on Move your practice and from the pages that hold each kind of record, often through an Import from a spreadsheet button.
- Choose your file and select Continue
- Check which of your columns go into which of our fields, then select Check the rows
- Look at the preview, tick any boxes the import asks for and select Import
- Read the result, and download the rows that did not import if there are any
Screenshot, for illustration only




Files the importer reads
- The spreadsheet import reads CSV files. If your list is in Excel or another spreadsheet program, save or export it as CSV first
- Encoding: UTF-8 (with or without the marker some spreadsheet programs add), then Windows-1252 and Latin-1, so accents survive a save from a desktop spreadsheet program
- Delimiter: commas, semicolons or tabs. The importer works out which your file uses, so a file saved with regional settings opens correctly
- The first row must hold the column headings, and no two columns can share the same heading. If two do, rename one and upload again
- Each kind of import has a limit on the number of rows in one file. The upload box shows it. A longer file can be split and uploaded in parts
Automatic matching
On the Match columns step, each column from your file is listed with the field it goes into. Columns that have the same name as one of our fields, or a name we recognise (for example, "E-mail" or "Date of birth (dd/mm/yyyy)"), are filled in for you. Under the column name a note says how the match was made and how sure we are, as Exact match or Recognised with a percentage.
- Capital letters, spaces, punctuation and anything in brackets are ignored when names are compared
- A column we cannot place is left as Don’t import this column until you choose a field for it
- Each of our fields can take only one column. If two of your columns look the same, one is matched and the other is left for you
Matching columns by hand
Use the drop-down beside any column to change where it goes, or choose Don’t import this column to leave it out. Fields the import needs are marked (required). While one of them has no column, a note says Map a column to that field, and Check the rows stays unavailable until you have done it. Each change is saved as you make it. If you want to begin again, choose Start again or Discard this import.
Saved mappings
When you change a column, your choice is remembered for your practice and for that kind of import. The next time you upload a file with the same column heading, your saved choice is used before automatic matching. In a firm, the saved choices belong to the firm, so a colleague uploading the same kind of list starts from the same matches. You can still change any of them on the Match columns step.
Suggest a mapping
When a required field still has no column, the Suggest a mapping button appears. It asks our own AI model, which runs on our servers, to look at the columns that were not matched.
- It sees the column headings and up to three sample values from each column, not your whole file
- It can only suggest fields that the import actually has
- Its suggestions fill the drop-downs and are marked Suggested by our AI. You can change any of them, and nothing is imported until you have checked the preview and chosen Import
- It does not use your AI credits
- If AI features are switched off for you or your firm, or the helper is not available just now, you will see a note and can match the columns yourself or ask our team to move your data for you
Lists that are not spreadsheets
Some lists only exist on paper or in a document. Every import takes a CSV, Excel, PDF or Word file as it is. For a photo of a paper list, on the Upload step, under What does your list look like?, choose A PDF, Word file or photo (read with our AI) instead of A spreadsheet or document (CSV, Excel, PDF or Word).
- Accepted files are PDF, Word (.docx) and photos (PNG, JPG or WebP), up to 10 MB each
- Our own AI model reads the document on our servers and turns it into rows. It does not use your AI credits
- The rows then go through the same Match columns step, the same preview and the same checks as a spreadsheet, and you decide whether to import them
- Please check the rows carefully. A reading of a document can contain mistakes, and the preview flags rows that do not pass the import’s own checks
- There is a daily fair-use limit on how many documents your practice can have read. If you reach it, you can upload a spreadsheet instead or ask our team to do it for you
- A price list for fees and services is read by that import’s own document reader instead
Asking an outside AI to read it again
After our own model has read a document, you can ask an outside AI to read the same document again for a second opinion. This is always your choice. The outside AI is a third-party provider, the same kind of provider the AI processing modes guide calls Cloud AI, but the second opinion asks for its own consent for each document, whatever your Cloud AI setting.
- Pick a model, then select Re-read with outside AI
- A window titled Send this document to an outside AI shows your AI credit balance and what the second opinion costs in credits
- Tick the consent box and choose Use outside AI. Your agreement is recorded, and the document is sent only after that
A second opinion uses AI credits from your balance, the same credits as other AI features. If you do not have enough, the window says so, and our own model and matching by hand are still available. The new reading replaces the rows from the first reading, and you go through the same steps again before anything is imported.
Checking before you import
The Check step is a dry run. It adds nothing, and shows counts for the whole file: rows, new, to fill in, already on file, differing from your records, and with problems. Below the counts are the first 50 rows with what would happen to each, and any problem found in a row, such as an email address that is not valid or a date that cannot be read.
Some imports ask you to confirm something before Import becomes available, for example a tick box or a choice of jurisdiction profile. Use Back to go and change the column matches.
Duplicate checks
Each row is compared with what is already on file, using the details that identify that kind of record (for clients, the email address, or otherwise name and date of birth). Every row gets one of four verdicts in the preview.
- Will be added: nothing on file matches, so a new record is created
- Will fill in missing details: one record matches, and the row adds details that record is missing. Values already on file are never overwritten
- Already on file, skipped: one record matches and the row adds nothing new
- Differs from what is on file, not changed: the row disagrees with the matching record, matches more than one record, or repeats an earlier row in the same file. These are left alone for you to sort out
Partial success and errors
If some rows have problems, the rest are still imported. When the import finishes, the result reads, for example, "12 imported, 3 filled in, 4 already on file, 1 left unchanged, 2 could not be imported".
If any rows could not be imported, choose Download failed rows. The file, import-errors.csv, lists each of those rows with its row number, your original columns, and a Problem column saying why. Rows left unchanged because they differ from what is on file are listed in it too. Correct those rows in your spreadsheet and upload it again. Rows that are already on file are skipped, so only the corrected rows are added. The file is written so that a spreadsheet program will not run any formula it contains.
How long a staged import is kept
Once you upload a file, its rows are held (staged) while you work through the steps. They are kept for 30 days. About 7 days before that, we email the person who uploaded the file to say it is about to expire, once. After 30 days the import is marked as expired and its staged rows are deleted, so you would upload the file again. You can also choose Discard this import at any point before importing, which removes the staged rows straight away. Discarding or expiry never removes anything that was already imported.
Can’t find what you’re looking for?
Our support team is here to help. Contact us and we’ll get back to you as soon as possible.
Contact us
