Point in Time Analysis

[darrellbutler]

Hi everyone, I'm looking for some advice please.

I have a KPI application which loads every night with fresh sales data, general ledger data etc. For such measures the application works fine and holds two years worth of historic data which day by day grows.

Some of the KPI's I need to measure look at stock levels, doing a complete refresh of data into my application every night will only show me the current situation.

What is the best technique to build up a history of point in time stock levels ?  possibly extract data daily into a QVD and build up a history ?Any help and advice welcome !

Regards

Darrell.

[Anonymous]

Hello

I work with monthly supply.
Every end of the month my application reads the table stock and stores the data in a QVD with the date of the month (MM/YYYY).
So I have a history of en my stock at the end of each month.

[#3]

I would create one new QVD per update and name them like; stock_level_today().qvd. No use loading all stock history every time you update, also if there is a error not all data will be lost. To ease use you can add a timestamp field into each qvd, that way you don't need to fetch it from the filename (and you can have it in any format you like).

If there is a need (load speed etc) I'd use a separate app to combine the files.

//Nils