You might want to merge segments to combine multiple history segments into one larger segment. For example, you can merge four different quarter segments into one segment that spans the entire year.

Before you can merge segments, create a segment set that contains the segments that you want to merge. For example, you might want to merge the segments "2015_01," "2015_02," "2015_03," and "2015_04" into one history segment that spans data for the year 2015. First, create a segment set that contains the four quarter segments. The name of the segment set must be unique and descriptive of the segment set.

After you create the segment set, run the merge partitions into single partition standalone job. When you run the merge partitions job, you select parameters such as the source connection name and the name of the segment set that you created. You can provide a tablespace for the merged segment, but if you do not, the job creates the segment in the existing tablespace of the most recent segment. If you provide a different tablespace, you must create the tablespace before you run the job.

After the job completes successfully, the older segments are merged into the most recent segment. In the

Manage Segmentation

window, the older segments have the status "Merged." The most recent segment is renamed to be a combination of the older segments. For instance, the most recent segment might be named "2015_01_2015_02_2015_03_2015_04." This segment has the status "Implemented."

If you change your mind about merging a segment set, you can replace the merged segment with the original segments. To revert the merge process, run the replace merged partitions with original partitions standalone job. For more on the replace merged partitions job, see the

Scheduling Jobs

chapter.

You can continue to create other segment sets and then run the merge job on those segment sets, provided that the segments exist in the same segmentation group. If you want to run the merge job multiple times for a segmentation group, mark the "FinalMergeJob" parameter as "N" until you run the final merge job for the segmentation group. On the final merge job for the segmentation group, set the "FinalMergeJob" parameter to Y.

You can repeat the merge process for each segmentation group in a data classification. You must complete all of the merge jobs for each segmentation group in a data classification before you run the "clean up after merged partitions" job. The clean up job drops the empty partitions and tablespaces, and merges the segment metadata. After you run the clean up job, the older individual segments are removed from the

Manage Segmentation

window and only the merged segment remains.

As a best practice, conduct a full backup before you merge segments.