overflow tables - manually inserts...

Author: dammie@seznam.cz (dammie)


I am storing long text (in column named comment) to table with overflow table (d_data and od_data)....

After store there is in table data string that ends with something as 001 ( that maybe means that another segment is 0001 in od_data)...

But what I dont understand is why before 001 is some char that gives no sense and seems be generated randomly! eg. #001,~001,A001

Any explanation? If I need to create data manually can I always use simply 0001 ?


Thanks for your opinions





  1. Hi David,

    NEVER insert/modify data in Tables with Overflow Tables
    (not even with variable length fields) outside of Uniface.

    The risk of loosing data is just too high.

    Only exception: Transfering data from one uniface controled table to another.

    Success, Uli




    Author: ulrich-merkel (ulrichmerkel@web.de)
  2. Hi Uli,


    but it is request of my customer. He send data that are 1500 char long I have to save them in table and they have to be accesible via uniface form :(


    Author: dammie (dammie@seznam.cz)
  3. What's the database? If you have control over the database, you can probably set the field type in Uniface to SC* and in SQL server it will be set to text (SQL server 2005, reset this to varchar(max).) Then there are no overflow tables, and thus no problem...

    Author: Iain Sharp (i.sharp@pcisystems.co.uk)
  4. As Iain has stated,

    there are some Interface-definitions which are mapped in the database to BLOBs, CLOBs, ....
    This way, you avoid segmentation and overflow tables.

    Have a look into the mapping descriptions of your uniface database driver.

    Success, Uli

    Author: ulrich-merkel (ulrichmerkel@web.de)
  5. Whats about writing a small uniface-programm which loads the customer data (I suppose a text file?) and let uniface write them?

    optional "Data conversion": Manipulate the customer-data to the correct xml-Format and import them with idf into your database.


    Author: gypsilon (wva@gypsilon.de)
  6. It is definitly not recommended, but also not impossible to manipulate overflow tables by hand.

    I remember for instance that we had to do that during a dataconversion a DELA.

    Maybe one of the DELA guys (Huub?) is reading this ....

    Author: Theo Neeskens (tneeskens@itblockz.nl)
  7. Hi David,

    I have never seen these random characters at the end of an overflow-field, but in my experience the final four characters always point to the next segment in the overflow table (so the final four characters in the comment field in the main table are 0001, the final four characters in the data field of the first row (segm = 0001) in the overflow table are 0002 and so on).

    I have used this to enter data using Cobol without problems.

    Hope this helps,



    PS. I agree with previous posts that in normal situations this is not a desirable way to work, and using Uniface is by far the safer (and easier) option.

    Author: Dennis van Duijn (dennis.van.duijn@sogeti.com)