Hi!

Hi!

I'm just starting to see some of the XE7 features and i have a question regarding the parallel library. Is it possible to use to loop through a dataset descent?

I have a project with dozens of loops through clientdatasets and it would be a major improvement for me if i could speed up those tasks.

Comments

  1. Multiple connections to different databases/tables in multithread modle is mostly much faster than one connection, just like Linas Naginionis  said.
     
    One thread per connection with one table/database is a good way to try and mostly thread-safe.

    ReplyDelete
  2. Sam Shaw Different databases/tables is quite a leap from the original post.

    ReplyDelete
  3. Thank you all for your feedbak. My ClientDatasets information is read from a file and live only in memory in the program execution. So i dont really have a database connection

    But there are certain tasks in wich a user changes a value in a grid that requires me to loop through 3 or 4 tables to validate or change data. And that needs to happen as fast as possible, so every milisecond that i could save would be great (to answer David Heffernan )

    I've seen the limitations that Marco Cantù was talking... My hope was that there was a way to access the cds lines as an array instead of using the built in cursors... but i guess not :P

    ReplyDelete

Post a Comment