Unconfigured Ad Widget

Collapse

Announcement

Collapse
No announcement yet.

Partitioning Files

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Partitioning Files

    A logical view created over the file or an SQL index will do the same thing. When you open the logical view instead of the physical file, the file is read by the key that was designated when the view was created.

  • #2
    Partitioning Files

    Depending upon circumstances, it might be possible to accomplish this with a multi-member physical file. Dave

    Comment


    • #3
      Partitioning Files

      You might be interested in Partitioned Tables

      Comment


      • #4
        Partitioning Files

        Thanks for your responses. I checked out the link and realized that partitioning on the iSeries is simply a modified multimember file but with some built in distributing to members. The biggest problem is that we have a windows application & a web application that will use these files. We are using Lansa so I don't know if their I/I modules support multimember files. I will have to contact them to find out. I learned something new today, so thank you all

        Comment


        • #5
          Partitioning Files

          Other than partitioning is there any other trick or technic that can be used to better performance of our SQL joins. We use SEQUEL to generate reports. We have an Order Detail file that currently has 1.7 million records. When we run different criteria and perform a join. SEQUEL calls the old optimizer and it doesn't always pick the best index.

          Comment


          • #6
            Partitioning Files

            I was recently approached with a question about partitioning a file. I believe this is a consept or ability on Mainframes and Unix systems. From what I was told basically a file is broken down into multiple components based on a set of specific critera. For example a 1 million record file can be partioned by date (lets say Month & Year). For all practical purposes programs see the file as 1 big file. Underneath the system has broken the file into a subset of the files. The depending on the criteria being used to access the file the system would only look at the partition it needs to. This is supposed to be a system type function. The database perspective is that there is 1 file. Can this type of partitioning of files be done on the iSeries (we are at 5V3)? This came up because of some performance issues with SEQUEL and potentially SQL in general. The joining of files in a select that the system does sometimes doesn't pick the correct access path (at least the one we think should be picked) and takes a long time to perform the join portion of the SEQUEL or SQL

            Comment


            • #7
              Partitioning Files

              You may want to read the following articles: 1. Performance Comparison of DDS-Defined Files and SQL-Defined Files; at: http://www.iseriesnetwork.com/resour...20067&channel= 2. Encoded Vector Index: at: http://www.itjungle.com/mgo/mgo050302-story01.html

              Comment

              Working...
              X