Tag Archives: Data Migration

Migrating Kenya Polytechnic to Koha ILS

Jan 21, 12
, , , , , , , , , , , ,
one comments

Busara Core Solutions has completed migrating Kenya Polytechnic University College -Library from Polymis to Koha. Although everything went down well as planned, it was all not an easy ride moving the old data to Koha. I encountered some challenges in the process of converting the data files from .ODB to MARC. Here is the brief:

When i got down to converting data from Kenya polytechnic to MARC format for the purpose of migrating to Koha, little did I know I had a huge challenge lying ahead. This is not to mean I expected an easy ride in the entire process, nope. In fact after reading a dozen documentations and internet posts and articles on data conversion, I knew the assignment was not a piece of cake.

Therefore, having not had any previous experience on dealing with data formats, I wanted to start with an application that would make it easier for me to understand what I was doing. So I avoided all terminal based applications (and there are quite many of them available on the Internet for free!)

I finally settled for MarcEdit version 5.5, a powerful C# programmed application made by some geniuses from Oregon (http://people.oregonstate.edu/~reeset/blog/ ). Although, am more comfortable with Linux based applications, installing MarcEdit on my Linux Mint – Julia laptop seemed an uphill task. The main reason being two dependencies which I was unable to resolve and especially: free MONO +1.2.6, a .NET framework required for Novell and Linux operating systems. Mid-way through my tireless efforts to get the dependencies resolved, I broke my dpkg backbone for debian, I gave up and switched to windows.

MarcEdit installed flawlessly on Windows, save for a few firewall issues which I resolved in no time by excempting all requests for MarcEdit.

Once MarcEdit was running, It was down to work. First (and easiest bit) was to convert the .mdb files to spreadsheet format. This I did quickly using OpenOffice Calc running on Windows 7. Then next was to organize the columns appropriately. My first obstacle was to concatenate the various fields which were spread out in the sheet to bring them close, at least something close to MARC format. After reading several articles (and thanks to google), I figure out the formula which came close to this = L1& “ ,”&K1& “ .”&M1.

Although I had to scratch my head several times (and almost lost my hair), once I had all the formula all figured out, I extremely enjoyed the entire process, but moving around over 40,000 records in a single file is no mean task. Then came the daunting Challenge: converting the file to MARC.

MarcEdit has an excellent delimited text translator, however, when I ran the translator on my .xls file, the editor crashed with a message that I could barely decipher:

After two to three days of goggling without a viable solution, I finally cracked it. And it was a rather simple solution. I converted my spreadsheet from .xls to .csv, loaded it into the delimited text translator and although translator exclaimed that it did not recognize the format, it loaded and mapped out all my fields like charm! Half of the task was done. I ran the MARC validator on the .mark file, then cleaned it up using Editplus and exported the record to .marc. The work was completed.

— Benson Mugambi