In the previous article, we extracted the SQL code from our old-style physical file into SQL tables. However, these tables are not much better than the original files. Let’s get to work on that and start to make them proper SQL tables.

So far in this series of articles, we have tables that are not really tables. First of all, we only generated their source code; we still need to run it to create the tables. Even if we do that, they are just siloed versions of the physical files they were generated from. In order to make them real tables, part of a real schema, we have a bit of work ahead of us.

CREATE TABLE Fundamentals

If everything went according to plan, you should see a new Run SQL Scripts window with the generated SQL code for the tables’ creation. Let’s analyze the PFSTM’s CREATE TABLE statement:

CREATE TABLE UMADB_CHP2.PFSTM (

-- SQL150B   10   REUSEDLT(*NO) in table PFSTM in UMADB_CHP2 ignored.

       STNM CHAR(60) CCSID 37 NOT NULL DEFAULT '' ,

       STDB DECIMAL(8, 0) NOT NULL DEFAULT 0 ,

       STAD CHAR(60) CCSID 37 NOT NULL DEFAULT '' ,

       STPN CHAR(15) CCSID 37 NOT NULL DEFAULT '' ,

       STMN CHAR(15) CCSID 37 NOT NULL DEFAULT '' ,

       STEM CHAR(60) CCSID 37 NOT NULL DEFAULT '' ,

       STDL CHAR(20) CCSID 37 NOT NULL DEFAULT '' ,

       STSN CHAR(11) CCSID 37 NOT NULL DEFAULT '' ,

       STSC CHAR(1) CCSID 37 NOT NULL DEFAULT '' )  

       RCDFMT PFSTMR     ;

Notice the column definition for STNM, for instance: It uses the DDS field definition (60A) to create a CHAR(60) column, but it adds a few things. It takes the CCSID from the physical file definition and adds the NOT NULL and a default. All the columns include the NOT NULL indication and have data-type–related defaults, mimicking DDS’s default behavior. This is exactly what can be expected, considering that corresponds to the “SQL version” of a DDS-defined file. If you look closely, you’ll see that the same record format name of the DDS file was extracted—essential for keeping RPG programs that use this table functioning without modification.

Also notice the comment line just below the first line of the statement: This is the database engine’s way of telling you, “Hey, I’m not DDS; there’s stuff I can’t do. Find a way to take care of this yourself.” In this case, it’s not a big deal, but there are situations in which you need to do some extra work to fully replicate a DDS-defined file. For instance, the EDTCDE keyword is not automatically converted, nor does it have an SQL equivalent. It’s true that it doesn’t directly impact the file itself, but if the field is referenced during a display or printer file creation, this keyword matters. There’s no silver bullet for this. You just have to be careful and analyze the impact of the changes.

If you run the statement as is, you’ll get an error because PFSTM already exists in schema UMADB_CHP2. So I’ll create schema UMADB_CHP3 and change the schema name in the CREATE TABLE statement to UMADB_CHP3. Now let’s run the statement. It should create the table in UMADB_CHP3, but you’ll see there’s something missing: All the descriptions are gone! If you scroll down a bit, you’ll see a couple of LABEL ON statements. Let’s change the schema name to UMADB_CHP3 in both statements, like this:

LABEL ON TABLE UMADB_CHP3.PFSTM

       IS 'UMADB - Students Master File' ;

LABEL ON COLUMN UMADB_CHP3.PFSTM

( STNM TEXT IS 'Student name' ,

       STDB TEXT IS 'Date of birth' ,

       STAD TEXT IS 'Home address' ,

       STPN TEXT IS 'Home phone number' ,

       STMN TEXT IS 'Mobile number' ,

       STEM TEXT IS 'Email address' ,

       STDL TEXT IS 'Driver's license' ,

       STSN TEXT IS 'Social security number' ,

       STSC TEXT IS 'Status' ) ;

Now let’s run these statements. Bada bing bada boom, the descriptions are back. Now seriously, this is just a reminder that, unlike DDS definitions, the CREATE and ALTER instructions don’t include descriptions. For that, you need to use a LABEL ON statement. It’s important to keep this in mind and make a habit of writing a LABEL ON statement right after every CREATE TABLE statement.

More to Come

At this point, we have crude but usable SQL tables. They are almost exact copies of our original, now inhabiting a new schema (UMADB_CHP3), that you can use in an RPG program. However, for non-RPG developers, these tables are kind of funny: short, cryptic column names, no keys whatsoever, nor default values for optional columns. This is what we’ll focus on next.