It’s time to finish the discussion about SQL’s CREATE TABLE instruction and provide a few more examples of what you can do with it!
I mentioned in previous TechTips that there were some additional tables in this imaginary database scenario. These tables provided definitions for the inventory master’s item, shelf, and warehouse ID columns. Let’s create these tables and introduce some additional features in the process. I’m going to start with the item master table:
CREATE TABLE MYSCHEMA.TBL_ITEM_MASTER
FOR SYSTEM NAME ITMMST
(ITEM_ID CHAR ( 15) NOT NULL WITH DEFAULT
ITEM_DESCRIPTION FOR COLUMN ITEM_DESC
VARCHAR ( 120),
ITEM_PICTURE FOR COLUMN ITEMPIC
BLOB ( 500K ))
So, what’s new here? I’m specifying the primary key directly in the ITEM_ID’s column definition. This is an alternative way of specifying a primary key, but you can use it only if your table has a single column as the primary key. The table’s next column is the item’s description, which is a VARCHAR data type with a length of 120 bytes. Note that this is only the maximum length, because the column might be left empty, in which case it will occupy zero bytes (null value) because I didn’t specify the NOT NULL option that you’ve seen in other columns—or have a description shorter than 120 characters.
Finally, the ITEM_PICTURE column is a Binary Large OBject (BLOB) occupying 500 Kb. The idea is providing a “storage area” for the item’s photo, directly in the database, thus facilitating the access to all item-related data when the table is accessed by an external application, like a web server. Even though this column’s content is not directly accessible in RPG, it can be manipulated in SQL, which you can encapsulate in an RPG procedure, as I explained in an earlier TechTip. Also, keep in mind that the BLOB column will occupy 500 Kb per row, regardless of the size of the file it contains. This behavior is the opposite of the ITEM_DESCRIPTION column’s, which will grow as needed up to a maximum of 120 bytes. This represents a slight deviation from the DDS Physical File to DDL Table conversion discussed in the previous TechTip, but it serves as a showcase of some of the features that SQL tables can provide. If I were going to use this table on an RPG program, I’d have to take into account the special characteristics of these two columns.
Having said that, we can move on to the juicy part! Now let’s create the warehouse master table:
CREATE TABLE MYSCHEMA.TBL_WAREHOUSE_MASTER
FOR SYSTEM NAME WHMST
(WAREHOUSE_ID FOR COLUMN WHID
DEC ( 8, 0) NOT NULL WITH DEFAULT
WAREHOUSE_NAME FOR COLUMN WHNAME
VARCHAR ( 80) NOT NULL WITH DEFAULT,
WAREHOUSE_ADDRESS FOR COLUMN WHADDR
VARCHAR ( 120) NOT NULL WITH DEFAULT,
WAREHOUSE_LATITUDE FOR COLUMN WHLAT
CHAR ( 15) NOT NULL WITH DEFAULT,
WAREHOUSE_LONGITUDE FOR COLUMN WHLON
CHAR ( 15) NOT NULL WITH DEFAULT,
CHECK (WAREHOUSE_ADDRESS <> ‘’))
This table’s definition is similar to the previously presented ones, but it includes a new concept: a constraint. Similar to its plain English meaning, an SQL constraint limits a column, in this particular case, the WAREHOUSE_ADDRESS column. The constraint defined here forces every insert and/or update to this table to provide a value, even if it’s just a blank space, to column WAREHOUSE_ADDRESS. You can create different types of constraints, such as forcing a column value to be unique in the table, which has a similar behavior to the DDS UNIQUE keyword, or the type of check you see in this example: allowing a change in a given column’s value if and only if a certain condition or set of conditions is met.
Now that I have created the item and warehouse master tables, I can complete my inventory master table definition. I know that the inventory master table depends on the former two tables, because any given item in the inventory master table must have valid item and warehouse IDs. If the inventory master table were a DDS-defined file, I’d probably write RPG code to enforce this relation, but InvMst is an SQL table, so there are other tools at my disposal. In the next TechTips I’ll introduce these tools, always building upon these tables, to keep a logical path for you to follow. In the meantime, explore and experiment with the CREATE TABLE statement and its potential. Share your doubts and ideas in the Comments section below!