iSeries, eServer i5, or System i5 System-to-PC Performance Considerations

Transferring data from the iSeries, eServer™ i5, or System i5® to the workstation depends on the following performance considerations:

The system workload.
How many records have to be looked at to complete the transfer.
If more than two files are joined. You need extra iSeries, eServer i5, or System i5 resources to join records from more than one file.
If GROUP BY fields are specified.
If complicated WHERE or HAVING comparisons are specified.

These factors and others influence the time needed to determine which data should be transferred. For example, the time needed to receive the first record of a transfer in which all the records are chosen is less than the time needed to start transferring a smaller group of records based on complicated WHERE or HAVING values. However, transferring all the records in a large file is sometimes impractical or unnecessary.

The iSeries-to-workstation data transfer function uses many functions within the iSeries, eServer i5, or System i5 to determine the fastest method of selectively retrieving records. When it selects a smaller group of records to transfer, the iSeries-to-workstation data transfer function uses the existing access paths whenever possible to improve performance.

For the iSeries-to-workstation data transfer function to consider using an existing access path (logical file), the access path must meet the following conditions:

It must be defined to the data that transfers.
It must have either *DELAY or *IMMED maintenance.

When you meet these conditions, you must then match the transfer request to the access path. The following considerations might be helpful when you define your transfer request:

The time it takes to select records based on WHERE clause values is less when the following things are true of the WHERE field:
- It is compared with a constant.
- It is the first key field in an existing access path defined to the data to be transferred.
A transfer request containing a GROUP BY or ORDER BY clause or both can work better if the key fields in the access path are in the same order as specified on the GROUP BY or ORDER BY clauses.
A transfer request containing a JOIN BY clause can work better when:
- An access path exists over the file that you are joining to.
- The field you are joining to is a primary key field in the access path.
- You are not returning records with missing fields.