From: dr-duplo <monsieur.cm@gmx.de>
Date: Mon, 3 Jun 2019 21:46:06 +0000 (+0200)
Subject: Concatenate HIKAZ data before MT940 parsing.
X-Git-Tag: v2.2.0~6^2
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=02e1be24eaf804f907e95b08a3d948cc750ed11e;p=thirdparty%2Fpython-fints.git

Concatenate HIKAZ data before MT940 parsing.

Some banks (especially Sparkasse) send HIKAZ data split over multiple segments, which led to MT940 parsing errors.
The patch changes the way the segments are parsed. Fixes #70 .
---

diff --git a/fints/client.py b/fints/client.py
index c5a7d72..855f732 100644
--- a/fints/client.py
+++ b/fints/client.py
@@ -498,12 +498,12 @@ class FinTS3Client:
             )
             logger.info('Fetching done.')
 
-        statement = []
-        for seg in responses:
-            # Note: MT940 messages are encoded in the S.W.I.F.T character set,
-            # which is a subset of ISO 8859. There are no character in it that
-            # differ between ISO 8859 variants, so we'll arbitrarily chose 8859-1.
-            statement += mt940_to_array(seg.statement_booked.decode('iso-8859-1'))
+        # Note 1: Some banks send the HIKAZ data in arbitrary splits.
+        # So better concatenate them before MT940 parsing.
+        # Note 2: MT940 messages are encoded in the S.W.I.F.T character set,
+        # which is a subset of ISO 8859. There are no character in it that
+        # differ between ISO 8859 variants, so we'll arbitrarily chose 8859-1.
+        statement = mt940_to_array(''.join([seg.statement_booked.decode('iso-8859-1') for seg in responses]))
 
         logger.debug('Statement: {}'.format(statement))