ibm midrange - Join Logical File on IBM i (AS/400) -

I see too many references for multi-format logical. What I want is something like this in the following one logical file? What kind of logical file is this? Here is the logical spec but it is not compiling.

In addition, if there is another modern way to do this, I would be happy to learn. These 2 files are PF whatever we want, it is a set of records from ICBLDTIR, which is the only place = pick, but this data is in the place master itself. Area LMLTPC

  R.PALREC JFILE (ICLOCMLM ICBLDTIR) CPD7989- * JOIN (ICLOCMLM ICTC1) IRLOC2 IRLOC3 LMLTPC S LMLTPC comp (choice of EQ)   
 Getting these errors:  
  * CPD7486 203 messages. . . . : Keywords specified at the wrong level * CPD7989 30 1 message. . . . The number of specifications is not valid    
 
 
 have you described in your question Is not that what we would say is a multi-format logical Each record your program will return the same format, called MYREC, even though that one format is composed of two files.  
 The solution to join a logical may have been recommended a decade or two ago, DDS is generally the direction you should focus on not improving performance and features with IBM SQL Focuses on SQL can give you more convenience, better overall performance, and make you more productive.  
 SQL provides you different types of methods that you want. The solution is to join a logical view which is closest to the logical file of your joining.  
  See Create MyView as select irloc1, irloc2, irloc3 to icbldtir iclocmlm at lmloc1 = irloc1 and lmloc2 = irloc2 and lmloc3 = Irloc3 and lmltpc = 'PICK'  < / Pre> 
 Or you can keep the SELECT statement in your program with embedded SQL.  
 To include it efficiently, you will want to make sure that you have a key access path (i.e., index) on each physical file (i.e. table) if it does not have three location fields, To SQL is quite simple to do.  
  create myFile index myfile_idx1 (field 1, field2, field3)   
 What do you say now? What we did was similar to a logical file that was essentially added. But you do not have the optimizer working on your behalf, but you can give it in hand In the current stage, with any method, the system will have to read a record from each file to create each result record. It must first read the index entry, then get the records specified by the index. But suppose we can actually read a record without actually reading it myself? This will speed things up and the system can usually fit more index entries on a memory page, an additional speed gain is achieved. So how do we do this? Well, the index includes all the information adapter needed, so it will use index-only access. So whatever we have to do, we put our additional area on the index. We do not need this as a field in a way, so we just add it as an additional area.   Create  iclocmlm_i2 at iclocmlm (lmloc1, lmloc2, lmloc3) RCDFMT iclocmlmr2 Add lmltpc