Parser and query API for bank CSV account transactions

bank, csv, transaction, transactions, money, sparebank, finance
pip install elv==1.0.14



Wheel Status

Elv is a small module for parsing bank transactions from CSV files and performing queries on the data.

Supported formats

It currently supports the banks who use the Eike Alliance solution. These include, but are not limited to

  • Jæren Sparebank
  • Sandnes Sparebank

but I'll add support for other banks whenever I get an example dump of their format. Send pull requests if you have them at

Sandnes Sparebank used to have another format, and I've kept it here, although I'm not sure anyone uses that anymore.

Norwegian short description

Elv er en Python modul for å lese banktransaksjoner eksportert fra banken din som en CSV-fil.

Foreløpig er det kun støtte for å lese filer fra Jæren Sparebank og Sandnes Sparebank. For å eksportere transaksjoner for en konto, velg et datointervall og deretter i øverste høyre ikon velger du "eksporter til fil".

Jeg har sett at andre nettbanker har tilsvarende funksjonalitet, så send meg gjerne eksempler på andre format så skal jeg legge dem til.


  • Parses CSV file from bank containing transactions
  • Money is stored in exact decimals
  • Contains a simple Python query API for sorting through large collections of transactions.
  • Can optionally put transactions in an in-memory SQLite3 database for even better queries.

Elevator pitch

Here's how easy it is to use elv:

>>> from datetime import date
>>> import elv
>>> trans = elv.parse("data.csv")
>>> trans
<Transactions:453 items from 2009-01-27 to 2015-03-20>
>>> trans[0].amount
>>> trans.range(date(2015,1,1), date(2015,3,1))
<Transactions:15 items from 2015-01-02 to 2015-02-20>
>>> trans.range(date(2015,1,1), date(2015,3,1)).total()


You can install from

$ python install # you may have to run as sudo

or from PyPI:

$ pip install elv

Example usage

If you have the bank account transactions in a file called data.csv, you can simply do:

$ python
>>> import elv
>>> transactions = elv.parse("data.csv")
>>> transactions
<Transactions:400 items from 2009-01-27 to 2014-09-29>
>>> transactions[0]
<Transaction:2014-09-29 2014-09-29 -2677.00  29519.13 'Vacation'>
>>> transactions[0].xfer, 9, 29)
>>> transactions[0].posted, 9, 29)
>>> transactions[0].amount

You can also get an in-memory SQLite3 database by doing:

>>> db = transactions.to_sqlite3()
>>> db
<sqlite3.Connection object at 0x10f31e200>
>>> db.execute("SELECT * FROM Transactions").next()
(0,, 9, 29),, 9, 29),
 u'Vacation', Decimal('-2677'), Decimal('29519.13'))

How to parse other formats

To parse the default format, Jæren Sparebank:

>>> ts = elv.parse("file.csv", format="Jæren Sparebank")

To parse Sandnes Sparebank:

>>> ts = elv.parse("file.txt", format="Sandnes Sparebank")

To see a list of formats, see the dictionary elv.formats.

The CSV File Format: Jæren Sparebank

The CSV file should be a plain text file with the ISO-8859-1 encoding (aka Latin1). It looks like this:

"31-12-2014";"31-12-2014";"Test 1";"-497,78";"5.520,09"
"30-12-2014";"31-12-2014";"Test 2";"-100,00";"6.017,87"
"30-12-2014";"31-12-2014";"Test 3 --æøåÆØÅ--";"-145,47";"6.117,87"
"30-12-2014";"30-12-2014";"Test 4";"-457,24";"6.263,34"
"29-12-2014";"29-12-2014";"Test 5";"-108,30";"6.720,58"

The fields are as follows:

  • Date when the transaction was placed, in format "DD-MM-YYYY".
  • Date when the transaction was posted ("bokført").
  • A message associated with the transaction, set by the one making the transaction.
  • The amount deposited or deducted from the account in NOK (Norwegian kroner) in format "-123,45", always two decimals and an optional sign.
  • The balance of your account after this transaction was completed.

The file itself contains no headers, and can contain many such transactions, one per line.


Copyright 2015, 2016 Christian Stigen Larsen

Licensed under Affero GPL v3 or later; see