⌘K

Data formats

How Mira reads the numbers and dates in a CSV: the organisation's default, and a format for one file.

Overview

Countries write numbers and dates in different ways, so the same text can mean different things:

Written in the fileRead the UK wayRead the Spanish way
1,005one thousand and fiveone point zero zero five
1.250,75not a numberone thousand two hundred and fifty point seven five

Dates have the same problem. 03/04/2026 is 3 April 2026 in the UK and Spain, but 4 March 2026 in the United States.

A CSV file doesn't say which way it was written, so Mira needs to be told. Your organisation sets a default that every CSV upload starts from, and any one file can be given its own format when it is uploaded.

Setting the default

The default is under Data formats on the General page of Settings. Open the organisation menu at the top of the sidebar, choose Settings, then General. Only Owners and Admins can change it. Each change saves straight away and applies to every upload from then on. A new organisation starts with 1,000.5 and DD-MM-YYYY.

SettingChoices
Thousands separatorComma 100,000, full stop 100.000, space 100 000, apostrophe 100'000
Decimal separatorFull stop 10.5, comma 10,5
Date and time formatThe order of day, month and year, the mark between them, and whether the time uses AM/PM. The full list is below.

The two separators can't be the same mark. If you choose the one the other already uses, they swap: setting the thousands separator to a full stop when the decimal is a full stop gives you 1.000,5, which is usually what you meant.

Common combinations

The thousands and decimal separators go in pairs. These are the ones you are most likely to meet.

Where it's commonThousandsDecimalA number
UK, Ireland, United States,.1,250.75
Spain, Germany, Italy.,1.250,75
Francespace,1 250,75
Switzerland'.1'250.75
Exports from databasesnone.1250.75

For a file with no thousands separator, like a database export, only the decimal separator has to be right.

Date and time formats

The format says which comes first, the day, the month or the year, and what separates them. The ones you are most likely to meet:

Where it's commonFormatA date
UK, Ireland, Spain, FranceDD/MM/YYYY25/03/2026
United StatesMM/DD/YYYY03/25/2026
Germany, SwitzerlandDD.MM.YYYY25.03.2026
Exports from databasesYYYY-MM-DD2026-03-25

Every format is a date followed by a time. Here is the full list, each showing 25 March 2026 at half past two in the afternoon.

OrderFormatExample
Day firstDD-MM-YYYY HH:MM:SS25-03-2026 14:30:00
DD/MM/YYYY HH:MM:SS25/03/2026 14:30:00
DD.MM.YYYY HH:MM:SS25.03.2026 14:30:00
DD. MM. YYYY HH:MM:SS25. 03. 2026 14:30:00
DD-MM-YY HH:MM:SS25-03-26 14:30:00
DD/MM/YY HH:MM:SS25/03/26 14:30:00
DD.MM.YY HH:MM:SS25.03.26 14:30:00
DD/MM/YYYY HH:MM:SS AM/PM25/03/2026 02:30:00 PM
Month firstMM/DD/YYYY HH:MM:SS03/25/2026 14:30:00
MM-DD-YYYY HH:MM:SS03-25-2026 14:30:00
MM/DD/YY HH:MM:SS03/25/26 14:30:00
MM/DD/YYYY HH:MM:SS AM/PM03/25/2026 02:30:00 PM
MM/DD/YY HH:MM:SS AM/PM03/25/26 02:30:00 PM
Year firstYYYY-MM-DD HH:MM:SS2026-03-25 14:30:00
YYYY/MM/DD HH:MM:SS2026/03/25 14:30:00
YYYY.MM.DD HH:MM:SS2026.03.25 14:30:00

One file in a different format

Sometimes one file doesn't match the rest, such as a report from a supplier abroad. You can set its format when you upload it, without changing the organisation's default. Anyone who can upload can do this.

  1. Add the file in the Add data window, as in Uploading files.
  2. At the end of the file's row, Mira shows the format it will read with as a sample, such as 1,000.5 · 25/03/2026. Click it.
  3. Choose the Thousands separator, Decimal separator and Date and time format this file uses. The sample in the row changes to match and turns darker, so you can see the file has its own format.
  4. Upload as normal.

To go back, open the same menu and choose Use the organisation default.

At Kestrel Outdoor the default is the UK format. A stock count arrives from the Hamburg warehouse with quantities like 1.250 and dates like 25.03.2026. Before uploading it, Priya sets that one file to a full stop for thousands, a comma for decimals and DD.MM.YYYY. Mira reads 1.250 as one thousand two hundred and fifty, not one and a quarter.

When a value doesn't match

If a column holds a value Mira can't read in the chosen format, it keeps the whole column as text rather than guess. A wrong guess would give you numbers that look right but aren't.

You can tell it happened when Mira won't add up or average a column you expected to be numbers, or won't sort dates in order. Upload the file again with the right format, or fix the stray values in the file first.

What data formats don't affect

  • Excel files. Mira reads numbers and dates straight from the workbook's cells, so the settings don't apply and Excel files have no format sample.
  • Connected systems. Data that comes in through a connector already has its numbers and dates typed, so nothing needs reading.
  • Held uploads. A file sent again from Held uploads is read with the organisation's default.
  • How numbers look in answers. These settings are about reading your files, not about how Mira writes numbers back to you.

Through the API

Files sent through Import data can carry their own format too. Add any of these form fields. Leave a field out and the organisation's default is used.

FieldAccepted values
NumericDivider, . ' or a single space
DecimalPlace. ,
DateTimeFormatAny format from the list above, written exactly as it appears there, such as DD/MM/YYYY HH:MM:SS
curl -X POST https://searchmira.com/api/v1.0/data/import \
  -H "X-API-Key: YOUR_API_KEY" \
  -F "ProjectID=YOUR_PROJECT_ID" \
  -F "NumericDivider=." \
  -F "DecimalPlace=," \
  -F "DateTimeFormat=DD.MM.YYYY HH:MM:SS" \
  -F "File=@/path/to/stock-count.csv;type=text/csv"

Send a space as exactly one space, not trimmed. A value Mira doesn't accept returns 400, and Msg lists the values it does.