A recent Cast Iron project involved taking tab-separated files and producing comma-separated files as a result. Seems easy enough? Of course... except three of the five use cases required merging multiple input files before producing the comma-separated output. Since merging files in memory is not a Cast Iron best practice, Rapid Integration used a MySQL database to stage the data. The database query activity allowed us to use SQL to aggregate, sort and dedupe the data so that Cast Iron could do what it does best.

Cast Iron does not like WSDLs that use SOAP encoding, so this project was a challenge. The input was a comma-separated file, and the output was a SOAP envelope containing the SOAP-encoded array of data required by the service we integrated. Given the nature of the array format, Rapid Integration decided that XSLT was the best option. So, the parsed flat-file becomes multiple parallel arrays in the SOAP envelope, and then we pass the result to the SOAP service. Since the documentation of the target endpoint was not very good, there was some trial and error over a day or so, but we got there in the end.