Ted Holt
-
Guru: Better Check Constraints
November 11, 2019 Ted Holt
This article has three purposes. If you use check constraints in your database, the purpose is to help you make better use of check constraints. If you don’t use check constraints, the purpose is to encourage you to use them and to point you in the right direction. If you already know all this stuff, the purpose is to goad you to email me and teach me something I don’t know.
The purpose of check constraints is to keep invalid data out of the database. That may seem unnecessary. Isn’t that what the applications are supposed to do? Yes, but …
Read more -
Guru: Remove Extra Blanks, Or Why I Attend Conferences
October 28, 2019 Ted Holt
Have you been to a conference lately? If not, you may be shortchanging yourself. I attend several conferences every year and I get immense benefit from them. I learn a lot, I get a break from the day-to-day, and best of all, I build relationships with other people.
I recently attended the RPG & DB2 Summit in Minneapolis, where I met a bright young developer named Kevan Robinson. He was kind enough to share his version of a tip that I shared with him and other attendees. It’s a technique that I learned ages ago from Craig Mullins, a mainframe …
Read more -
Guru: Debugging Common Table Expressions
October 21, 2019 Ted Holt
I cannot say enough good things about common table expressions. Words like wonderful and marvelous don’t begin to describe them. However, CTEs do add a bit of complexity to an SQL query, and when the result set doesn’t contain the correct results, any common table expression can be the culprit. Fortunately, debugging queries with common table expressions is not difficult.
To illustrate what I mean, let’s assume we have a query that retrieves shipment information for one day. It involves a few tables:
- a one-row table containing a shipment date
- a table of shipment header information
- a customer master table
-
Guru: MERGE, Chicken, And Eggs
September 30, 2019 Ted Holt
Which came first: the chicken or the egg? I don’t have time to ponder such trivialities. However, I am glad to know that SQL has a way to help me with chicken-and-egg database updates, i.e., when two statements need to run but each politely needs for the other to go first.
Suppose you support an IBM i system that keeps up with inventory. It has an item master table (physical file) that stores general information such as a description, the standard cost, and the list price of an item.
create table ItemMaster (ItemNumber char(6), Revision dec(3), Description char(20), Cost dec(5,2),
… Read more -
Guru: Enumerated Data Types In RPG
September 16, 2019 Ted Holt
IT has changed a lot since I entered the field several decades ago, but some things have not changed. I would read in those early days that COBOL was dead, and I read the same thing now. Yet COBOL is 60 years old and still going strong. Back then I heard RPG criticized as “Real Poor Garbage”. These days I hear it scorned as “legacy”, which I assume is supposed to mean the same thing. Yet today’s RPG is better than any of its predecessors for business programming.
RPG supposedly does not have the features of modern languages. Maybe not, …
Read more -
Guru: I’m A Number, You’re A Number, Everybody’s A Number
September 9, 2019 Ted Holt
Do you, like Bob Seger, sometimes feel that you are nothing more than a number? Me too. That’s because to many people, that’s exactly what we are. And if there’s one thing that computers are good at, it’s assigning numbers — to orders, to accounts, to invoices, to transactions, and of course, to people. Since we have to make the computer assign numbers, we may as well learn the modern way to do it.
In my earliest days of programming, I would store the last assigned of a series of numbers in a data file. (The S/34 and S/36 …
Read more -
Guru: Read a Data Area As a One-Row Table with SQL, Take Two
August 26, 2019 Ted Holt
Fifteen years ago, reader W.G. asked me about the possibility of treating a data area as a one-row table (a physical file with one record) in an SQL query. The question intrigued me because in my System/36 days, I had often wished that I could access the local data area (LDA) as a one-record data file in a query.
Today, thanks to Scott Forstie and his team at IBM, I update my response to W.G. with more information. It’s not that the technique I presented in 2004 is outdated — it’s as relevant as ever — but that the fine …
Read more -
Guru: One Way To Deal With Two Null Formats
July 8, 2019 Ted Holt
Hey, Ted:
We are building a new system and want to use modern programming and database techniques. I have had quite a time trying to get nulls to act right. It gets confusing fast because RPG handles them differently than the way embedded SQL does. When using SQL for I/O, how do we handle the two null formats?
— Brian
There are several ways to go about the “problem” of nulls. Let me give you one simple method, but keep in mind that it’s not the only way.
First, let’s create a table and put some data into it.
create
… Read more -
Guru: When %SCAN Isn’t Sufficient
June 24, 2019 Ted Holt
The RPG %SCAN built-in function is wonderful! I can still remember having to look for a string within a string using RPG II on System/36. What an ordeal that was! Yet in some situations %SCAN can’t do all I need it to do. In those cases, I rely on the power of SQL.
One case where SQL comes in handy is when I need a case-insensitive scan. Instead of RPG’s %SCAN function, I use SQL’s LOCATE and UPPER functions, like this:
dcl-s Description char(48); dcl-s pos int (10); exec sql set :pos = locate ('HAMMER', upper(:Description));If Description has the …
Read more -
Guru: Use SQL to Generate Random Data
June 17, 2019 Ted Holt
Suppose I needed to generate a large database table with random data in order to adequately test the performance of an SQL query. Suppose that, for security reasons, I was not allowed to copy the production version of the table. Suppose that I needed a way to generate a lot — and I do mean a lot! — of random data. Suppose this scenario is not mere supposition.
Before an SQL query goes into production, it should be tested against a production-like dataset. Running a query against a test dataset of 25 rows (records) can produce unpleasant surprises when it’s …
Read more