> ## Documentation Index
> Fetch the complete documentation index at: https://orgo.space/docs/llms.txt
> Use this file to discover all available pages before exploring further.
# Preparing your file
> Get the spreadsheet right before you upload it: CSV format, the 50 MB limit, headers, dates, yes/no columns and choices
Most import problems are decided in the spreadsheet, long before Orgo sees it. This
page is what to fix while the file is still open in Excel, Numbers or Google Sheets.
**Built for** the person holding an export from an old system, a membership register
kept in a spreadsheet, or a list somebody else prepared. **Replaces** the guesswork of
uploading, reading the errors, and trying again.
***
## Save it as CSV
Orgo reads **CSV** and **TXT** files. It does not read Excel workbooks. An `.xlsx`
or `.xls` file has to be saved as CSV first, and so does an `.ods` or a Numbers
document.
| Where the file lives now | How to get a CSV |
| --------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------- |
| **Excel** | File, Save As (or Export), and pick **CSV UTF-8 (Comma delimited) (.csv)**. Choose that entry rather than plain "CSV" when both are offered. |
| **Numbers** | File, Export To, **CSV**, and set **Text Encoding** to **Unicode (UTF-8)**. |
| **Google Sheets** | File, Download, **Comma Separated Values (.csv)**. Google Sheets always writes UTF-8. |
| **Another system's export** | Ask for CSV. If the export only offers Excel, open it and save it again as CSV. |
Save one sheet per file. A workbook with several tabs becomes a CSV of whichever
sheet was active, and the rest is lost silently.
Save as **UTF-8** wherever the option exists. Orgo also recognises the older Western
European encodings that Excel produces on Windows, but UTF-8 is the only one that
carries accented names, Cyrillic and Greek without any guessing. A UTF-8 byte order
mark at the start of the file is harmless and is removed for you.
***
## Keep it under 50 MB
The limit is **50 MB per file**. The size is checked in your browser before the
upload starts, so an oversized file never leaves your computer: you get **"File is
larger than the 50 MB import limit"** straight away.
A CSV of member records is small, so 50 MB is a lot of rows. If you do go past it,
split the file:
1. Open the spreadsheet and split it by row into two or more files, for example
rows 1 to 40,000 and rows 40,001 onwards.
2. Copy the header line into every part. Each file has to carry its own headers.
3. Upload the parts one after another, mapping each one as its own import.
Splitting is also the faster route for very large files, not only the legal one.
A file of a few tens of thousands of rows moves through Orgo noticeably quicker
as several parts than as one enormous upload.
***
## How Orgo reads the file
**One header line, then one row per record.** No merged cells, no blank spacer
rows in the middle, no totals row at the bottom, no notes written above the
headers. A spreadsheet meant to be read by a person is rarely a good import file;
strip the presentation out of it.
**The header line is found for you.** Blank lines above the header are skipped, and
when an import starts Orgo looks through the first ten lines for the row that
carries the columns that type needs. What you should still do is put the headers on
the first line with content in it, because the mapping screen builds the sheet you
see from that line. A title such as "Members export 2026" sitting above the headers
will be read as the header row there.
**The delimiter is detected, not configured.** Orgo counts semicolons, commas and
tabs on the header line and uses whichever appears most often. That is why the
header line matters more than the rest of the file: a header full of punctuation
can outvote the real separator. Values that contain the delimiter are fine as long
as your spreadsheet quoted them, which every export listed above does.
**One row is one record.** Five rows for the same person are five rows, not one
person seen five times, and each of them is processed on its own.
***
## Column headers are yours to name
This is the change that matters most if you have imported into Orgo before.
**Your headers no longer have to match Orgo's field names.** You are not renaming
`Email address` to `email`, or `Date of birth` to `birth_date`, before you upload
anything. You upload the file as your old system exported it, and then you say on
screen which column goes where. `E-mail`, `Correo`, `Adresa de email` and
`Contact email` are all fine, because you point each one at the destination field
yourself in the mapping screen.
Two things follow from that:
* **Do not rename headers to guess at Orgo's vocabulary.** Leave them as they are.
Clear, human headers are easier to map than half-remembered technical ones.
* **Every column needs a header.** A column with an empty header cell cannot be
mapped, so give it a name, even a rough one, or delete the column.
Headers that do already match Orgo's names are still worth having: the mapping
screen recognises them and pre-selects the destination for you, which leaves less
to do by hand. It is a shortcut, not a requirement.
Two columns whose headers differ only by capitals, spaces or hyphens count as the
same column, so `First Name` and `first-name` collide and only one survives. Give
every column a distinct name.
***
## Values that need care
Most cells are copied across as text and need no thought. These are the ones that
quietly do something other than what the spreadsheet looks like it says.
| Column holds | Write it as | What happens otherwise |
| --------------------------- | --------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **A date** | `2026-03-14`, always year, month, day with dashes | Anything else is refused: `14/03/2026`, `2026-3-14`, `March 14 2026`, and the plain numbers Excel leaves behind when a date column loses its formatting. A date Orgo cannot read fails the whole row, not just that cell. |
| **Yes or no** | `true`, `1` or `yes` for yes, in any capitalisation | Only those three spellings count as yes. Everything else, including `y`, `x`, `da`, `ja`, `oui`, `sim` and a tick character, is stored as **no**, with no warning and no error. |
| **A number** | The number as you want it stored | Numbers are stored exactly as written. Nothing is parsed, rounded or reformatted, so `12,50`, `1 250` and `EUR 50` all land as that text. Decide on one way of writing amounts and apply it to the whole column. |
| **Nothing (an empty cell)** | Leave it empty only when you mean "no change" | An empty cell writes nothing at all. It cannot clear a value that is already stored, so blanking a column in the spreadsheet and importing again leaves the old values in place. |
| **One choice from a list** | The option label as it reads in Orgo | The value is matched against the existing options by label, ignoring capitals and extra spaces, so `chess`, `Chess` and `Chess ` all find the same option. A value that matches nothing creates a new option on the field. |
| **Several choices at once** | The labels separated by one consistent character | Multi-choice columns are split on a delimiter you choose in the mapping screen, and each part is then matched like a single choice. Mixing separators inside one column leaves you with options named after the whole cell. |
The yes/no rule is the one that catches people out. A column of `da` and `nu`, or
`ja` and `nein`, or `Y` and `N`, imports as a column of no's, and the import
reports a clean run because nothing went wrong from its point of view. Convert
those columns to `yes` and `no`, or to `1` and `0`, in the spreadsheet first. A
find-and-replace on the column takes a minute and is the single highest-value
thing on this page.
Only capitals and spacing are forgiven when a choice is matched. `Chess` and
`Chess.` are two different options, and so are `Football` and `Fotbal`. Tidy the
spelling in the column before you import, or you will find near-duplicate options
on the field afterwards.
***
## Start from an example file
Every import type on the **Data Import** page carries an **Example CSV** link on its
step card, next to the upload button. The example is a real, working file for that
type with a handful of demonstration rows.
Use one as a template:
1. Download the example for the type you are importing.
2. Open it and look at the columns. They are the ones that type understands, written
the way Orgo writes them.
3. Replace the demonstration rows with your own data, keeping the header line.
4. Save as CSV and upload.
Working this way means your headers already match Orgo's names, so the mapping
screen has almost nothing left to ask you. It is also the quickest way to see what a
type expects when the field list is unfamiliar. Any column of your own that the
example does not have can simply be added: you will map it when you upload.
***
## Before you upload
* [ ] The file is saved as **CSV or TXT**, not as an Excel workbook, and in UTF-8.
* [ ] It is under **50 MB**, or it has been split into parts that each carry the header line.
* [ ] The **headers sit on the first line with content**, with nothing above them.
* [ ] **Every column has a header**, and no two headers are the same.
* [ ] There is **one row per record**: no merged cells, no spacer rows, no totals line.
* [ ] Every date column reads **`2026-03-14`**, in that order, for every row including the last ones.
* [ ] Every yes/no column uses **`yes`, `1` or `true`**, and nothing else means yes.
* [ ] Nothing you wanted to erase was left blank, because a blank cell changes nothing.
* [ ] Files about people carry an **email column** wherever you have addresses.
* [ ] You scrolled to the **bottom of the file** and checked that the last rows look like the first ones.
That last one is worth the thirty seconds. Exports often start clean and turn messy
several thousand rows in, where a different date format or a stray note was pasted
years ago.
***
## Troubleshooting
The limit is 50 MB and it is checked in the browser, so the message appears
before anything is uploaded. Split the spreadsheet by row into two or more
files, copy the header line into each part, and import the parts one after
another. If the file is unexpectedly large for its row count, look for empty
columns stretching far to the right, or thousands of empty rows below the data:
delete them and export again.
Orgo accepts one date format, `2026-03-14`. Everything else, including
`14/03/2026` and `14.03.2026`, is refused. In the spreadsheet, select the date
column and give it a custom format of `yyyy-mm-dd`, then export again and open
the CSV in a text editor to confirm the dates really were written that way.
Watch for the case where Excel shows a date but has stored a serial number: the
CSV then contains something like `46095`, which is not a date to Orgo.
Only `true`, `1` and `yes` are read as yes. Any other word, in any language, is
stored as no without an error, so the import looks successful. Fix the column in
the spreadsheet with a find-and-replace, `da` to `yes` and `nu` to `no` for
example, then import the corrected file. Because a blank cell changes nothing,
re-importing the same file with the column corrected will set the values that
should be yes.
Orgo needs one. Add a line at the top naming each column. The names are yours to
choose, since you map them on screen, so plain words are enough. Also check that
nothing sits above that line: a report title, a date stamp or an exported note
on line one will be treated as the headers by the mapping screen. Delete those
rows and save again.
The separator is chosen by counting semicolons, commas and tabs on the header
line. Two things break it. First, punctuation inside the header names
themselves, such as a column called `Name, first`: rename it. Second, a file
that mixes separators, usually because it was edited by hand after export. The
reliable fix is to open the file in your spreadsheet, check that the columns
split correctly there, and export it once more so the whole file uses one
separator. In Excel on a computer set to a European locale, prefer the
**CSV UTF-8** export, which is consistent about what it writes.
***
## Related
* [Data import](/docs/platform/customization/import) - the whole import, from upload to finished records
* [Mapping columns](/docs/platform/customization/import/mapping-columns) - the screen where your headers meet Orgo's fields
* [Custom fields in an import](/docs/platform/customization/import/custom-fields) - storing columns Orgo has no field for
* [Linking records](/docs/platform/customization/import/linking-records) - how a row finds the member, event or product it belongs to
* [Import reference](/docs/platform/customization/import/reference) - the fields each import type reads
* [Custom Fields](/docs/platform/users/custom-fields) - the profile fields an import can fill or create