About data compression in Containers
You can optionally use Data compression to save a summary of the information in a container when there is not enough room to hold every original data point. When compression is added to the container, compression then occurs automatically, reducing multiple data items within a container to a single data item, based on defined compression parameters and the compression fields in the Type Descriptor selected on the container details screen.
When compression is used, the Type Descriptor that is used in the Container must be defined with compression parameters.
The compression process uses the field designated as the GroupBy to group multiple data items into a single data item, for those with the same GroupBy field value. The compression process then sums the values of the fields that are defined with the compression parameter.
The following table provides examples of the timing of compression and the results that occur.
Period Start
Time Unit
Period Length
Compression Results
0
Day
1
Compression occurs, immediately, during the run as the container is accessed. The data stored in the container that day is compressed.
For example, on Monday, data is stored in the container. That data is compressed on Monday, at the completion of the transaction set, for a given ID.
1
Day
1
Compression occurs each day, during the run as the container is accessed. The data stored in the container the previous day is compressed.
For example, on Monday, data is stored in the container. When the container is accessed during the Tuesday run, the data that was stored on Monday is compressed. Tuesday’s data remains untouched.
7
Day
1
Compression occurs each day, during the run as the container is accessed.
For example, if a user is making many transactions every week. By compression the information daily, you can for example determine the maximum, sum, or the average daily transaction for this customer over the course of a week.
2
Week
1
Compression occurs every two weeks, at the first run following a two-week cycle. The data stored in the container during the first week of the previous two-week cycle is compressed. Compression occurs as the container is accessed.
For example, this month Sundays fall on the 1st, 8th, 15th, 22nd, and 29th. The last time the container was compressed was on the 1st. You are now doing a run for the 15th. The first time the container is accessed during the run, all items dated from the 1st through the 14th are compressed. The date - timestamp of record is the same as the uncompressed record that is closest to the 14th.
Compression will occur when you do the run for the 29th. During this run the records date - timestamped from the 15th through the 28th are compressed.
1
Month
1
Compression occurs every month, at the first run following a one-month cycle. The data stored in the container during the previous month is compressed. Compression occurs as the container is accessed.
Copyright IBM Corporation 2012. All Rights Reserved.
