I generate invoice as PDFs or files of various types, eg XML, CSV. I had trouble retaining the original line sequence as per the source transaction. Typically I scan a data source of invoice headers, and within the scan I filter another data source for all the invoice lines for that invoice, then scan those lines to generate the output: FOR EACH <invoice header data source> - Using HFilter, filter the invoice lines data source for that invoice (ordered by invoice and linenumber) - FOR EACH <invoice lines data source> - - Print the invoice report or create the output file I tried 3 different ways to filter the invoice lines data: 1. FOR...EACH <invoice lines data source> - did not return the correct set of data, even though a display of the data source in the debugger showed the correct set of lines. 2. WHILE HOut(<invoice lines data source>) - correct invoice set but the lines were in random order 3. FOR EACH <invoice lines data source> where "saleid = '" + <invoice number data source>.saleid + "'" ON LineNumber #3 worked correctly. The "where clause" created the correct subset of data, and "ON linenumber" retained the correct line sequence. See [doc.windev.com]. This may be well known but not to me. Shared for anyone who can benefit! https://www.wxforum.info/read.php?27161,799185,799185#msg-799185 Tue, 19 May 2026 02:31:52 +0200 Phorum 5.2.23 https://www.wxforum.info/read.php?27161,799185,799250#msg-799250 https://www.wxforum.info/read.php?27161,799185,799250#msg-799250 Mike James WinDev Solutions Sat, 02 May 2026 00:23:41 +0200 https://www.wxforum.info/read.php?27161,799185,799244#msg-799244 https://www.wxforum.info/read.php?27161,799185,799244#msg-799244
Over the years I have found Windev functions that work perfectly and those that don't as I am sure you have.
One function that randomly causes issues for me is the For Each construct, especially if you do some process inside the loop so now I never use it.

For an invoice, I load all the line records into a memory table with an incrementing numeric counter field, then sort the table on the counter field and use it to supply the body section of the invoice report.

Regards
Al]]> Al WinDev Solutions Thu, 30 Apr 2026 02:32:27 +0200 https://www.wxforum.info/read.php?27161,799185,799243#msg-799243 https://www.wxforum.info/read.php?27161,799185,799243#msg-799243
"ON LineNumber" (a field name from the table/data source) replaced by "ON sFieldName" (a string containing the fieldname).

It's not the first time we've had to make a similar change.]]> Mike James WinDev Solutions Thu, 30 Apr 2026 02:19:29 +0200 https://www.wxforum.info/read.php?27161,799185,799185#msg-799185 https://www.wxforum.info/read.php?27161,799185,799185#msg-799185
Typically I scan a data source of invoice headers, and within the scan I filter another data source for all the invoice lines for that invoice, then scan those lines to generate the output:

FOR EACH <invoice header data source>
- Using HFilter, filter the invoice lines data source for that invoice (ordered by invoice and linenumber)
- FOR EACH <invoice lines data source>
- - Print the invoice report or create the output file

I tried 3 different ways to filter the invoice lines data:

1. FOR...EACH <invoice lines data source> - did not return the correct set of data, even though a display of the data source in the debugger showed the correct set of lines.

2. WHILE HOut(<invoice lines data source>) - correct invoice set but the lines were in random order

3. FOR EACH <invoice lines data source> where "saleid = '" + <invoice number data source>.saleid + "'" ON LineNumber

#3 worked correctly. The "where clause" created the correct subset of data, and "ON linenumber" retained the correct line sequence.

See [doc.windev.com].

This may be well known but not to me. Shared for anyone who can benefit!]]> Mike James WinDev Solutions Thu, 19 Mar 2026 23:52:08 +0100