Reducing the amount of unnecessary Index Advisor data

I do check the Index Advisor on a regular basis to see what is recommended. I have found that this perspective can become overwhelmed by old data. Who cares what index were advised over a year ago? I either addressed that advice, or the program(s) were changed that used the files in that way.

Every six months I go into the tool and clear the data. This way I will only see the recent recommendations. I have an entry in my calendar to remind me to do this. Wouldn't it be nice if this could be automated.

As part of the latest Technology Refreshes, IBM i 7.6 TR1 and 7.5 TR7 a new Global variable is introduced that will automate the purging for me.

Limiting the amount of data SELF retains

While I love the amount of information that the IBM i will retain for us, there are times where it can become overwhelming, or redundant. An example of this is how much SELF, SQL Error Logging Facility, data do I wish to retain.

Prior to the latest Technology Refreshes, IBM i 7.6 TR1 and 7.5 TR7, I was responsible for purging the SELF error log table, QSYS2/SQL_ERRORT. As part of these TRs a new Global variable was introduced.

Finding Table functions with SYSFUNCS

In every IBM i recent release and Technology Refresh a few new Audit Journal Table extract functions are added. I wanted to see which one were on the IBM i partition I was using. I was tempted to use the OBJECT_STATISTICS Table function to get a list of them:

01 SELECT OBJNAME,OBJLONGNAME 02 FROM TABLE(QSYS2.OBJECT_STATISTICS('SYSTOOLS','*ALL','AUD*'))

Line 1: I am only interested in the system object name, OBJNAME, and the long (SQL) object name, OBJLONGNAME.

Line 2: The Audit Journal Table functions are in the SYSTOOLS library. I am using the wild card for the object name, in the third parameter, to find all the objects that start with 'AUD'.

The results were:

Retrieving the description for the Audit Journal Entry Types

I have, to date, been disappointed that I could not programmatically retrieve a description for the 420 Audit journal entry types. Referring to IBM's documentation was just not practical. A new SQL View in the Fall 2025 Technical Refreshes has given me what I wanted, a list of all the journal entry types with their descriptions.

The View is JOURNAL_CODE_INFO, which is found in the library QSYS2. It contains the following columns:

1. JOURNAL_CODE:  The journal code
2. JOURNAL_CODE_DESCRIPTION:  Description of the Journal code
3. JOURNAL_ENTRY_TYPE:  Journal entry type
4. JOURNAL_ENTRY_TYPE_DESCRIPTION:  Description for the journal entry type

The journal codes are for a specific "action", while the Journal Entry Type is the type of information. You can see this with the new View with the following statement:

March presentations

I will be presenting at two events this month:

The first is the March meeting of the Central Texas IBM i User Group, CTXiUG, on Tuesday March 3, 6:30 PM (CT). It is online and free to everyone.

You can learn more about the meeting and register here.

The following week I am off to my first conference of 2026: WMCPA's iCon 2026 conference, in Lake Delavan, WI, Tuesday March 10 - Thursday March 12. Attending this event is a great way to kick-off the conference season. WMCPA always make this a wonderful conference, full of great presenters and subject matter.

You can learn more about it, who will be speaking and what about, and register here.

Don't forget to ask for your RPGPGM.COM-unity ribbon when you see me.

I hope to see you there!

Improvements to GENERATE_SPREADSHEET

The Fall 2025 Technical Refreshes brought some significant improvements to the GENERATE_SPREADSHEET SQL scalar function. These being:

• Execute SQL statement within an IFS files
• Append results from another SQL statement to an existing spreadsheet
• Create a new worksheet, and start the output at a specific row and column
• End the Java program that is started to generate the spreadsheet
• No longer have to use STDOUT to capture information from GENERATE_SPREADSHEET

I am not going to repeat what I have said in previous posts. But, I am going to give an example, below of generating a spreadsheet as a reminder:

IBM Bob formally announced

On Friday February 20 IBM's AI Bob was formally announced with a planned availability date of March 24.

You can read the announcement here.

IBM document: IBM i 7.5 vs 7.6 comparison

On January 30, IBM Support published a page called "IBM i V7R5 versus V7R6 comparison" (shame on IBM Support for not using the proper names of 7.5 and 7.6).

The page is to answer the question: What are the differences between 7.5 and 7.6?

I thought it was something useful to share because, as well as the comparison, there are links to many useful pages of information. You can reach the page here

I hope you find it as interesting as I did.

Find all the services programs used by a program

Most of the time it is my curiosity, but there have been occasions, where I needed to know if a program calls a particular service program. It is easy to discover which service a programs call a service program, but what about the service programs that that those service programs calls? Now there is way to do that.

A new SQL Table function, PROGRAM_RESOLVED_ACTIVATIONS, has been added to IBM i as part of latest Technology Refreshes, IBM i 7.6 TR1 and IBM i 7.5 TR7.

PROGRAM_RESOLVED_ACTIVATIONS has five parameters:

1. PROGRAM_LIBRARY:  Library that contains the object
2. PROGRAM_NAME:  Name of the program or service program
3. OBJECT_TYPE:  *PGM if the object is a program, *SRVPGM if it is a service program
4. DEFERRED_SERVICE_PROGRAMS:  NO do not follow deferred activations, these are service programs that had *DEFER in the BNDSRVPGM parameter of the CRTPGM or CRTSRVPGM command. YES follow the deferred activations, this is the default
5. IGNORE_ERRORS:  NO if an error is found an error happens. YES a warning is returned; this is the default

I have only used the first three.

Using SQL TRANSLATE in place of REGEXP_REPLACE

I encountered something last week that I think may be a "bug" in either RPG or in Db2 for i. I was using the REGEXP_REPLACE scalar function, and I received an error message that made no sense. I ran out of ideas of how to overcome this error using REGEXP_REPLACE, and I went on to use the TRANSLATE scalar function. Before I tell you what the error is I need to explain my situation.

I was working upon a partition which was running IBM i 7.5 with the latest PTFs for Technology Refresh 7, TR7, and the latest PTFs for Database and RPG:

• Database (SQL) = SF99950 Level 11
• RPG = 5770WDS SJ08064

The partition's CCSID is:

01 SELECT SYSTEM_VALUE_NAME,SYSTEM_VALUE 02 FROM QSYS2.SYSTEM_VALUE_INFO 03 WHERE SYSTEM_VALUE_NAME = 'QCCSID' SYSTEM_VALUE_NAME SYSTEM_VALUE ----------------- ------------ QCCSID 273

2026 Fortra survey published

This year's Fortra 2026 IBM i Marketplace Survey results has been published. This is the twelfth year of the survey, and I for one, always find it an interesting read to learn what the other members of the IBM i community are up to.

If you did not participate in this year's survey I would like to strongly encourage you to participate in the next, as it is a good insight as to what is important to people and companies using IBM i.

You can read it here.

You can also watch the discussion about the survey with Steve Will, IBM i CTO and Chief Architect, Tom Huntington, Fortra, Kris Whitney, IBM i Chief Engineer, Tomothy Prickett Morgan, IT Jungle, here.

SEND_MESSAGE can now send message to any message queue

When SEND_MESSAGE SQL procedure was introduced in 2021, as part of IBM i 7.3 TR10 and 7.4 TR4, it would be used to send messages to the System operator message queue. I have used this many times to do so, but I have always wondered how I could do the same to any message queue? Well, in the latest Technology Refreshes, IBM i 7.6 TR1 and 7.5 TR7, this functionality was added.

Two new parameters were added to the procedure to make this possible:

• MESSAGE_QUEUE_LIBRARY:  The library that contains the message queue I want to send the message to
• MESSAGE_QUEUE:  The message queue I want to send the message to

If I omit these parameters the message will be sent to the System operator message queue.

The first example I am going to show is using the procedure as it was intended, sending a message to the System operator message queue. I am going to use the message id CPF9898, which is a general escape message, i.e. I can provide any text I want for the message to use. I can show this using the MESSAGE_FILE_DATA SQL View: