Problems with Using Pervasive Rebuild Utility on APOPNFIL and AROPNFIL Tables

Q - Recently, we used Pervasive's rebuild utility to rebuild all Elliott database tables (i.e., all \Elliott7\DATA\*.BTR files).  After the rebuild, we noticed some strange things happening to our A/P and A/R modules.  For example, below is a sample of our A/R aging report for customer "01CSV." As you can see, invoice number 352545 should be the first record on the report, followed by the payment and credit memos afterward.  But as you can see we now have payments above and below the invoice record. This is confusing when we try to research accounts and resolve disputes with the customer.  Can this be corrected?

A - We think the problem is caused by Pervasive's rebuild utility. Generally speaking, you can use the Pervasive rebuild utility for all Elliott database tables. There are two exceptions: 

  • APOPNFIL.BTR: This is the A/P Open Item table. Rebuilding it with the  Pervasive rebuild utility can cause major problems with AP payment processing. 
  • AROPNFIL.BTR: This is the A/R Open Item table. Rebuilding this can cause minor inconveniences with A/R account inquiry or aging report printing. 

For AROPNFIL.BTR, there are two keys -- the primary key and the alternate key. The alternate key has two parts: 

  • Customer Number
  • Apply-To No. 

When the A/R aging report is printed by the Apply-To order, systems use the alternate key. As you can see in your example, customer # “01CVS “ has multiple AR Open Item records for the same Apply To “352545.” When the first key value of “01CVS “ + “352545” (i.e., the invoice record) was being written to the AROPNFIL.BTR table, there was no issue. When the second record of the same key value was being written, PSQL handled the duplicate alternate key value by creating a “link list.”  The first record is on the first node of the link list, the second record is on the second node, and so on and so forth by the chronological order of when the record was added. Therefore, when you print the A/R Aging Report, the record of the same apply-to value will appear in the same sequence. 

When you perform a rebuild through the Pervasive utility, unfortunately the original link list sequence of the duplicate key value is not guaranteed. Therefore, it shows up as the problem you reported. This is not a critical problem because the A/R aging report still works correctly other than the minor printing sequence issue. If you don’t like this, you can follow the procedure below to fix it:
(1) At the command prompt, go to your Elliott DATA folder, and export AROPNFIL.BTR to AROPNFIL.TXT:
(2) At the command prompt, rename AROPNFIL.BTR to AROPNFIL.OTR
(3) At the command prompt, clone an empty AROPNFIL.BTR from AROPNFIL.OTR.
(4) At the command prompt, load AROPNFIL.TXT to AROPNFIL.BTR
Now go print the A/R aging report again. It should print by the proper date sequence as you expected.

Why did the above procedure fix your problem? When you use the BUTIL -SAVE option, it exports AROPNFIL.BTR to AROPNFIL.TXT by the primary key sequence. AROPNFIL.BTR primary key consists of:

  • Customer Number
  • Document Date
  • Doc Number
  • Document Type
  • Apply To Number 

Therefore, the data in AROPNFIL.TXT is sorted by customer number first, then document date sequence second. When you use BUTIL-LOAD option to load AROPNFIL.TXT to AROPNFIL.BTR, it will write the alternate key by document date sequence within the same customer record, so the link list of the duplicate key value will be re-established correctly.

Of course, you should only apply the above procedure when all Elliott users are out of the system.

You should apply the same procedure for the APOPNFIL.BTR table. We think the strange issues with the AP module are related to the same problem. The AP module tends to behave badly when the APOPNFIL.BTR duplicate alternate key original link list sequence is disturbed. Therefore, we ask our users never use the Pervasive rebuild utility on APOPNFIL.BTR or APHISFIL.BTR (this is the history version of APOPNFIL with identical table structure). In your case, you might want to skip AROPNFIL.BTR and ARHISFIL.BTR as well.

Feedback and Knowledge Base