Skip to main content

CSV

A CSV file has a line per record, its fields split by a delimiter, usually with a header line naming them:

Payment ID,Date,Amount,Sort code
P-1,2026-03-02,100.5,040004

Tell Crosswalk how to read it​

To read a file, Crosswalk needs its schema: the name and type of each field, and which ones are lists. The builder below makes one from an example, and shows the JSON it reads as you go. When it reads the example the way you want, the Schema tab has the schema to use in a map.

The builder below starts with a file of payments. What it reads is underneath, and changes as you go.

  1. Click a field on any line to select its column, then name it and set its type in the bar under the lines. A column of numbers with leading zeros, such as a sort code, starts as text, so the zeros aren't lost.
  2. Check What it reads: that's the JSON a map gets.
  3. To do the same for your own file, open Sample and paste a few lines of it.
Sample 3 lines, CSV: Payment ID,Date,Amount,Sort code

Paste a few lines of your own file to make a schema for it; the layout is worked out from what you paste.

Lines
Payment ID,Date,Amount,Sort code
P-1,2026-03-02,100.5,040004
P-2,2026-03-03,50,040004
Click a field or a part of one to name it. Shift-click another field to select the fields between, to make a list of them.
Writes the sample back as it was.
{
"rows": [
{
"paymentId": "P-1",
"date": "2026-03-02",
"amount": 100.5,
"sortCode": "040004"
},
{
"paymentId": "P-2",
"date": "2026-03-03",
"amount": 50,
"sortCode": "040004"
}
]
}

A file with several kinds of line, each starting with a tag, such as a header, details and a trailer, works too: see several kinds of line. Every setting is in the flat files reference.