Sample session for incremental data transfer

This tutorial describes how to transfer data to an external file, change the data in the table, and then use the transfer command again to repopulate the table from this external file, showing that transfer appends the data to the file, and doesn’t overwrite it.

NoteAlthough this example transfers data out of, and back into, the same table, in a typical user scenario, data would be transferred out of one table and then into a different one.

  1. Create the spt_TableTransfer table, which stores transfer history:

    sp_setup_table_transfer
    
  2. Configure max transfer history. The default is 10, which means that Adaptive Server retains 10 successful and 10 unsuccessful transfers for each table marked for incremental transfer. This example changes the value of max transfer history from 10 to 5:

    sp_configure 'max transfer history', 5
    
    
    Parameter Name               Default     Memory Used
    	Config Value    Run Value              Unit
    	Type                      Instance Name                  
    -------------------------     -------     -----------
    --------------   --------------------   -------
    ---------------          ---------------
    max transfer history      10              0
    5                5                        bytes               
    dynamic                  NULL                           
    
    
  3. Create the transfer_example table, which has the transfer attribute enabled and uses datarow locking:

    create table transfer_example (
    f1 int,
    f2 varchar(30),
    f3 bigdatetime,
    primary key (f1)
    ) lock datarows
       with transfer table on
    
  4. Populate the transfer_example table with sample data:

    set nocount on
    declare @i int, @vc varchar(1024), @bdt bigdatetime
    select @i = 1
    while @i <= 10
    begin
       select @vc = replicate(char(64 + @i), 3 * @i)
       select @bdt = current_bigdatetime()
       insert into transfer_example values ( @i, @vc, @bdt )
       select @i = @i + 1
    end
    set nocount off
    

    The script produces this data:

    select * from transfer_example
    order by f1
    
    f1       f2                             f3
    --------- ------------------------------ ----------------------------
       1      AAA                            Jul 17 2009  4:40:14.465789PM 
       2      BBBBBB                         Jul 17 2009  4:40:14.488003PM 
       3      CCCCCCCCC                      Jul 17 2009  4:40:14.511749PM 
       4      DDDDDDDDDDDD                   Jul 17 2009  4:40:14.536653PM 
       5      EEEEEEEEEEEEEEE                Jul 17 2009  4:40:14.559480PM 
       6      FFFFFFFFFFFFFFFFFF             Jul 17 2009  4:40:14.583400PM 
       7      GGGGGGGGGGGGGGGGGGGGG          Jul 17 2009  4:40:14.607196PM 
       8      HHHHHHHHHHHHHHHHHHHHHHHH       Jul 17 2009 4:40:14.632152PM 
       9      IIIIIIIIIIIIIIIIIIIIIIIIIII   Jul 17 2009 4:40:14.655184PM 
      10      JJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ  Jul 17 2009 4:40:14.678938PM
    
  5. Transfer the transfer_example data to an external file using the for ase format:

    transfer table transfer_example
    to 'transfer_example-data.ase'
    for ase
    
    
    
    (10 rows affected)
    

    The data transfer creates this history record in spt_TableTransfer:

    select id, sequence_id, end_code, ts_floor, ts_ceiling, row_count
    from spt_TableTransfer
    where id = object_id('transfer_example')
    
    id        sequence_id      end_code     ts_floor       ts_ceiling 
    row_count            
    ---------   -----------       --------      -------------   -------------------
    592002109  1                0            0              5309	                    10
    
  6. Disable the transfer attribute from transfer_example to demonstrate that the receiving table does not need the transfer attribute enabled to receive incremental data (the database must have select into enabled before you can run alter table):

    alter table transfer_example
    set transfer table off
    

    After the alter table command runs, spt_TableTransfer is empty:

    select id, sequence_id, end_code, ts_floor, ts_ceiling, row_count
    from spt_TableTransfer
    where id = object_id('transfer_example')
    
    id        sequence_id      end_code     ts_floor       ts_ceiling 
    row_count            
    ---------   -----------       --------      -------------   -------------------
    
    (0 rows affected
    
  7. Update transfer_example to set its character data to no data and to specify a date and time in its bigdatetime column so you can verify the table does not contain the original data.:

    update transfer_example
    set f2 = 'no data',
    f3 = 'Jan 1, 1900 12:00:00.000001AM'
    
    
    
    (10 rows affected)
    

    After the update, transfer_example contains this data.

    select * from transfer_example
    order by f1
    
    f1         f2                              f3                            
    ----------- ------------------------------ ---------------------------
    1           no data                        Jan  1 1900 12:00:00.000001AM
    2           no data                        Jan  1 1900 12:00:00.000001AM 
    3           no data                        Jan  1 1900 12:00:00.000001AM 
    4           no data                        Jan  1 1900 12:00:00.000001AM 
    5           no data                        Jan  1 1900 12:00:00.000001AM 
    6           no data                        Jan  1 1900 12:00:00.000001AM 
    7           no data                        Jan  1 1900 12:00:00.000001AM 
    8           no data                        Jan  1 1900 12:00:00.000001AM 
    9           no data                        Jan  1 1900 12:00:00.000001AM 
    10          no data                        Jan  1 1900 12:00:00.000001AM 
    
    (10 rows affected)
    
  8. Transfer the example data from the external file into transfer_example. Even though transfer_example is no longer marked for incremental transfer, you can transfer data in to the table. Because it has a unique primary index, the incoming rows replace the existing data and do not create duplicate key errors:

    transfer table transfer_example
    from 'transfer_example-data.ase'
    for ase
    
    
    (10 rows affected)
    
  9. Select all data from transfer_example to verify that the incoming data replaced the changed data. The transfer replaced the contents of transfer_example.f2 and transfer_example.f3 tables with the data originally created for them, which were stored in the transfer_example-data.ase output file.

    select * from transfer_example
    order by f1
    
    f1      f2                             f3                            
    ------- ------------------------------ ----------------------------
       1    AAA                            Jul 17 2009 4:40:14.465789PM 
       2    BBBBBB                         Jul 17 2009 4:40:14.488003PM 
       3    CCCCCCCCC                      Jul 17 2009 4:40:14.511749PM 
       4    DDDDDDDDDDDD                   Jul 17 2009 4:40:14.536653PM 
       5    EEEEEEEEEEEEEEE                Jul 17 2009 4:40:14.559480PM 
       6    FFFFFFFFFFFFFFFFFF             Jul 17 2009 4:40:14.583400PM 
       7    GGGGGGGGGGGGGGGGGGGGG          Jul 17 2009 4:40:14.607196PM 
       8    HHHHHHHHHHHHHHHHHHHHHHHH       Jul 17 2009 4:40:14.632152PM 
       9    IIIIIIIIIIIIIIIIIIIIIIIIIII    Jul 17 2009 4:40:14.655184PM 
      10    JJJJJJJJJJJJJJJJJJJJJJJJJJJJJJ  Jul 17 2009 4:40:14.678938PM 
    
  10. Reenable transfer for transfer_example so subsequent transfers use, by default, the previous parameters:

    alter table transfer_example
    set transfer table on
    
    (10 rows affected)