How arrays are used in synon?

Answer 1

Synon array processing is totally intuitive for any synon user. You do not need to open a manual. If you can use a file, you can use an array. The main difference is that there are only 4 function types: CRTOBJ, CHGOBJ, DLTOBJ, RTVOBJ, and only 1 "ACP" (set of Key fields)

The array processing really behaves more like an indexed file than an array. The file portion (Records) is a multiple occurrence data structure and the index portion is an array. The data structure contains the fields as defined in the *ARRAY file. The array(Index) is the fields defined as the Key concatenated with the occurrence# of the data structure. When you create an element in the array (a record in the file), the program takes the next available occurrence of the data structure and loads the fields. The program also creates the Index element by concatenating the key fields with the occurrence# of the newly added data structure. The program then does a Sort Array to sequence the Index so that the records may be accessed in key sequence. Any time a record is deleted, the Index element is blanked out and the Index is Sorted (Sort Array). This is a very clean solution and the performance is surprisingly good considering that the array is sorted after each add or delete.

When accessing the data, a LOKUP is performed on the array using the key value with blanks instead of the occurrence# and looking for GT in the array. This allows processing of partial keys as positioners or restrictors. The program then reads the data structure for the occurrence# taken from the array element to get the data(fields). For the next occurrence, it is not necessary to do another LOKUP (which is slow) because the array is sorted. The program only has to read the next element in the array and continue the processing as long as the partial key matches the key in the array. From the point of view of the Synon user, this processing is identical to a file RTVOBJ.

Synon loads the array from the bottom (an RPG technique to improve performance in large arrays) because the LOKUP command may be given a starting position, but cannot be stopped until every element is checked when no match is found. By loading the array backwards from the bottom and counting the elements, the search may be started at the last element loaded and only elements which contain data are read.IF No Match is Found

An array is defined for 1000 elements.

Only 10 elements are loaded.

These are loaded to occurrences 991 thru 1000.

When the LOKUP is performed, it starts at 991 and only 10 compares are made.

** If the array is loaded from the top, 1000 compares are performed.