Where problems are encountered with the size of the paging area, you should specify the maximum number of documents to be read. This avoids the risk of the program crashing. If you specify a maximum number, the system reads documents only (and not the master data and transaction figures). If the latter two items are to be read, they must be selected in a subsequent run (i.e. without specifying a maximum number). Example: You have a large number of documents which have to be processed in several stages. You therefore select the documents according to their posting date, e.g. from 01.01.1996 to 31.12.1996. You also enter a maximum number e.g. 500,000. The program then extracts around 500,000 line items. If the maximum number of documents is reached, the last document number read is output. The next time your run the program, you enter this document number as the lower limit. If all the documents in the given time period are processed, when the program is later run their master data is read (with no entry as to maximum number of documents). For this run, you then enter an "X" as document number since no more documents need to be read. The advantage of this procedure over selecting the documents according to the CPU or posting date is that you do not need to know how the documents are spread across the time period in question. In one transfer, for example, all documents are posted under a CPU date. Furthermore, transaction figures are not selected until the last run (and not every run). Note: Choose the value 500,000 if your paging area is 512 MB.