== Why is this an issue?

The use of a result data structure with file I/O improves performance because it moves the data in one large block from file to data structure (or vice versa) rather than field by field. 


Additionally using a data structure can limit the problems caused by having bad data in a file. Without a data structure, the entire ``++READ++`` operation will fail at the first bad value. With one, the error comes only when the bad field is used.


=== Noncompliant code example

[source,rpg]
----
     F MyFile   IF   E             Disk
       * Noncompliant
     C                   READ      Record1
----


=== Compliant solution

[source,rpg]
----
     F MyFile   IF   E             Disk

     D InputData       DS                  LikeRec(Record1)
     C                   READ      Record1       InputData
----


ifdef::env-github,rspecator-view[]

'''
== Implementation Specification
(visible only on this page)

=== Message

Use a result data structure for this file I/O.


'''
== Comments And Links
(visible only on this page)

=== on 2 Apr 2015, 17:42:59 Ann Campbell wrote:
http://www.bmeyers.net/faqs/14-tips/32-rpg-iv-style?start=3

http://www.ibmsystemsmag.com/ibmi/developer/rpg/Ending-Those-Decimal-Data-Error-Blues/?page=1

endif::env-github,rspecator-view[]