Some Kanji sequences in QR Codes scan incorrectly

Discussion created by jbante on Jun 10, 2016
Latest reply on May 9, 2017 by TSGal

Product and version: FileMaker Go 15

OS and version: iOS 9.3.2 (13F69)

Hardware: iPad Air 1



Using the Insert from Device script step to scan a QR Code, certain sequences of Kanji characters get scanned as different, often shorter, sequences. I confirmed that the QR Codes are encoded correctly by scanning them using the Scandit app. QR Codes containing only one character at a time for each character in any of these sequences scan correctly, and QR Codes containing the same characters in a different order may scan correctly. I first noticed this in QR Codes where the data is encoded only using QR Code's Kanji mode, but the exact same behavior holds when the same data is encoded using UTF-8 mode instead. I've found two particular patterns:


1. Triplets that scan incorrectly as a group, but all subsequences scan correctly, the first and last character alone scans correctly, and triplets made by swapping the first and last characters also result in a QR Codes that scan correctly.

2. In the process of looking for examples of (1), I also found cases where a pair using the first and last characters of the triplet also does not scan correctly.


How to replicate


See the sample QR Codes in the attached file.


Workaround (if any)