I’m having a fantastic morning already. I’m not really in that great of a mood but the major upgrade that was performed to the applications I’ve been working on took place last night. There was a lot riding on it. I previously was using a disconnected dataset to perform lookups and matching to load order specific comments into another dataset that held parts dynamically allocated to specific orders. This added an overhead of about four minutes to the load time of the application to individually do this for each application when it first started.
I knew right away something had to be done about this. I’ve batched everything else into processes that are fired off during the initial parts pegging application, and this seemed like another need area. SO after identifying all the possible solutions to this problem on paper I went about trying each. Finally I found the solution. Create a new field in the database where the parts were stored to, and then fill in the comments each morning. The programs would run as they normally do with the user totally oblivious to where the comments were stored. At the end of the night the batch processes would run, wipe out the part list, regenerate the part list, and while doing so would do the dynamic lookup of previously issued comments using the new part criterion.
Last night it was made apparent this was needed to be pushed out into production. Everything was relying on this program, and thankfully it worked like a charm. However, it wasn’t all roses and chocolates. The program’s normal run time has been steadily increasing, due to the MS Access connection limitations, and due to the networked file share location of the database. The program has to update the database by copying over 15,000 parts, 35,000 build dates, updating the initial 50,000 parts looped through, and maintaining data on about 45,000 PO and WO combinations. Needless to say each one takes a few milliseconds to complete, but the sheer amount of data bogs down Access to where it takes several hours to update all this information.
So my task today, while keeping mind on the other applications I’m still working on is to speed up this process even more. It’s kind of hard using the outdated technology that I’m using, the older .Net framework, and the limitation to using MS Access. So we’ll see what I can cook up. Either way the new process worked on its first time in the sandbox. Hopefully we can make it peppy now. Stability and reliability come first right?
Business