


The WHEN directive is used when you want to include multiple record definitions or REDEFINES in the XFD for use with your Windows application. It's typically used when you want to force certain columns of data to be available that wouldn't be available otherwise.
Recall that the key fields and the fields from the largest record are automatically included as explicit columns in the "virtual" database table. So you should use the WHEN directive if you want the user to be able to access all the data in the COBOL file in a way that is understandable.
WHEN declares that the field (or subordinate fields, if it is a group item) that immediately follow the directive must appear as a column (or columns) in the "virtual" database table. It also states one condition under which the columns are to be used. WHEN thus guarantees that the fields will be explicitly included in the table (as long as they aren't FILLER and don't occupy the same area as key fields).
Syntax
$XFD WHEN field operator value
or
*(( XFD WHEN field operator value ))
Field is the name of a data item that corresponds to a field. If there is an XFD NAME directive for this data item, the name used in the XFD WHEN directive is the name given to the item by the XFD NAME directive, not its COBOL name.
The operator specifics the relation between the field value and the alphanumeric literal that satisfies the condition. Operator can be one of the following:
| Operator | Relation |
| = | The field value is equal to the literal value. |
| != | The field value is not equal to the literal value. |
| > | The field value is greater than the literal value. |
| < | The field value is less than the literal value. |
| >= | The field value is greater than or equal to the literal value |
| <= | The field value is less than or equal to the literal value |
$XFD WHEN field = other
"Other" is true only when all other conditions for the same field are false. For example, if your FD contains the following lines of code
$XFD WHEN ATYPE = "C"
$XFD WHEN ATYPE = "D"
$XFD WHEN ATYPE = other
the "other" condition holds true only if both "atype = 'c'" and "atype = 'd'" are false.
Example
The following code demonstrates an example of using the WHEN directive.
01 key-record.
* employee-number is a key data item
03 employee-number pic 99999.
03 emp-type pic x.
$xfd when emp-type="1"
* record has this form when emp-type=1
01 data-record-1.
03 filler pic 99999.
03 filler pic x.
03 name-1 pic x(35).
03 pay-rate-1 pic 99.99.
$xfd when emp-type="2"
* record has this form when emp-type=2
01 data-record-2
03 filler pic 99999.
03 filler pic x.
03 name-2 pic x(35).
03 pay-rate-2 pic 999.99.
03 subordinates pic 999.
03 position pic x(50).
The effect of these directives is to force "emp-type", "name-1", and "pay-rate-1" to correspond to fields, even though they are not in the largest record description. Therefore, the corresponding table has the following fields:
EMPLOYEE_NUMBER
TYPE
NAME_1
PAY_RATE_1
NAME_2
PAY_RATE_2
SUBORDINATES
POSITION
If each data item is subordinate to at most one XFD WHEN directive, as in this example, for such a data item:
If a data item is subordinate to two or more XFD WHEN directives, the following applies:
and the database table in the usual way.
WHEN Directive With TABLENAME Clause
A WHEN condition can also be assigned a tablename, so that AcuODBC will consider the different conditions as separate tables. If you assign a tablename, the data that immediately follows the WHEN directive and meets the specified condition will be considered as a separate table. Therefore, you will have the tables specified by the WHEN directive plus the tables specified by the TABLENAME clause.
If you select one of these conditions as a table, the columns of that table will consist of all columns that depend on that condition. Note that if a condition is not named, that condition will not have a table associated with it.
When an XFD names a condition, such as WHEN, AcuODBC presents multiple tables from a single XFD file. The tables include the current name of the file, as well as any named conditions, with a tablename parameter listed as the name of the table (see syntax section below).
You cannot include a HIDDEN field in a WHEN directive with a TABLENAME clause. This is due to the complexities of editing or adding records. With a hidden field, you must add the data, but since the field is hidden, you cannot see it and you might add a value that would cause unexpected results.
If an XFD file in an alias contains WHEN directives with TABLENAME phrases, the corresponding tables are defined in the usual way, using the data file specified by its physical file name. You can define two or more file aliases with the same XFD file but with different physical file names if the XFD file does not contain any WHEN directives with the TABLENAME clause.
Syntax
$XFD WHEN field operator value TABLENAME=name_of_new_table or *(( XFD WHEN field operator value TABLENAME=name_of_new_table ))
The syntax is essentially the same as for the WHEN directive alone, with the addition of the TABLENAME clause The word OTHER can be used only with "=". It means "use the following field(s) only if none of the other WHEN condition(s) listed for the same field is met." In other words, this condition is true only if all other conditions for the same field are false.
For example:
.
.
assign to "ar_table"
.
.
01 ar-code-type.
*(( xfd when ar-code-type = "s" tablename=ship ))
03 ship-code-record pic x(4).
*(( xfd when ar-code-type = "b" tablename=backorder ))
03 backorder-code-record redefines
ship-code-record.
*(( xfd when ar-code-type = other ))
03 obsolete-code-record redefines
ship-code-record.
If you tried to connect to this data source through a program like Access, you would see three tables: ship, backorder, and ar_table. If you placed $XFD COMMENT ACUODBC READ-ONLY TABLE immediately before the "xfd when ar-code-type = "s" tablename=ship" line, then the ship table and ar_table would be read-only, but the backorder table would not.
OTHER may be used before one record definition, and may be used once at each level within each record definition.
If the following code were compiled without directives, the underlined fields would appear explicitly in the database table. Note that the key fields are included automatically, as are the fields from the largest record. FILLER would be ignored:
01 ar-codes-record.
03 ar-codes-key.
05 ar-code-type pic x.
05 ar-code-num pic 999.
01 ship-code-record.
03 filler pic x(4).
03 ship-instruct pic x(15).
01 terms-code-record.
03 filler pic x(4).
03 terms-rate-1 pic s9v999.
03 terms-days-1 pic 9(3).
03 terms-rate-2 pic s9v999.
03 terms-descript pic x(15).
If you added the WHEN directive as shown below, it would cause the fields from the SHIP-CODE-RECORD to be included in the database table, and would determine when specific database columns would be used. The underlined fields would appear as columns in the database table:
01 ar-codes-record.
03 ar-codes-key.
05 ar-code-type pic x.
05 ar-code-num pic 999.
$xfd when ar-code-type = "s"
01 ship-code-record.
03 filler pic x(4).
03 ship-instruct pic x(15).
$xfd when ar-code-type = "t"
01 terms-code-record.
03 filler pic x(4).
03 terms-rate-1 pic s9v999.
03 terms-days-1 pic 9(3).
03 terms-rate-2 pic s9v999.
03 terms-descript pic x(15).
FILLER data items don't have unique names and are thus not used to form columns in the database table. You could use the NAME directive to give them a name if you really need to see them in the database table. However, in this example the FILLER data items implicitly redefine key fields. Thus, they would be disregarded even if you provided a name for them.
Example 2
In the following code, in which no WHEN directives are used, the underlined fields will be explicitly named in the database table. (Key fields have the suffix "key" in their names in this example.)
Note that REDEFINES records simply re-map the same data area and are not explicitly included in the database table by default:
01 archive-record.
03 filler pic x(33).
03 archive-code pic 9(6).
03 archive-location pic 9(2).
03 filler pic x(10).
01 master-record.
03 animal-id-key.
05 patient-id pic 9(6).
05 species-code-type pic 9(5).
05 species-name pic x(6).
03 service-code-key.
05 service-code-type pic 9(6).
05 service-name pic x(10).
03 billing-code.
05 billing-code-type pic 9(4).
05 plan-name pic x(8).
03 office-info.
05 date-in-office pic 9(8).
05 served-by-name pic x(10).
03 remote-info redefines office-info.
05 van-id pic 9(4).
05 proc-code pic 9(8).
05 vet-name pic x(6).
If you added the WHEN directives shown below, you would add several columns to the database table. The fields that would appear in the table are underlined:
*(( xfd when animal-id-key = "00000000000000000" ))
01 archive-record.
03 filler pic x(33).
03 archive-code pic 9(6).
03 archive-location pic 9(2).
03 filler pic x(10).
*(( xfd when animal-id-key = other ))
01 master-record.
*(( xfd use group ))
03 animal-id-key.
05 patient-id pic 9(6).
05 species-code-type pic 9(5).
05 species-name pic x(6).
03 service-code-key.
05 service-code-type pic 9(6).
05 service-name pic x(10).
03 billing-code.
05 billing-code-type pic 9(4).
05 plan-name pic x(8).
*(( xfd when billing-code-type = "1440" ))
03 office-info.
05 date-in-office pic 9(8).
05 served-by-name pic x(10).
*(( xfd when billing-code-type = other ))
03 remote-info redefines office-info.
05 van-id pic 9(4).
05 proc-code pic 9(8).
05 vet-name pic x(6).
Example 3
If your application has a REDEFINES whose field names are more meaningful than the fields they redefine, you might consider switching the order of your code, rather than using a WHEN directive. Use the less significant field names in the REDEFINES.
For example, you might change this:
03 code-info.
05 filler pic 9(8).
05 code-1 pic x(10).
03 patient-info redefines code-info.
05 patient-id pic 9(4).
05 service-code pic 9(8).
05 server-name pic x(6).
to this:
03 patient-info.
05 patient-id pic 9(4).
05 service-code pic 9(8).
05 server-name pic x(6).
03 code-info redefines patient-info.
05 filler pic 9(8).
05 code-1 pic x(10).
The fields that would appear in the database table by default are underlined above. This shows how the column names might become more meaningful when the order is reversed. Your application operates the same either way.
This COBOL code:
01 col-type pic x.
03 col-def.
$xfd when col-type = "a"
05 def1 pic x(2).
$xfd when col-type = "b"
05 def2 redefines def1 pic 9(2).
results in this database table:
| col_type | def1 | def2 |
| a | xx | null |
| b | null | 10 |
| a | yy | null |