DataMystic

This error occurs when a line to be processed does not end with the specified delimiter,
and when the filter is Move fields: #-delimited field END-1..END-0 move to 1

The error is a show-stopper, as it interrupts the processing of the rest of the input file.

In the attached filter, suitable lines are in the trial run area.

I have included examples with both # and comma as delimiters, with lines both with & without the final delimiter.

It matters not what kind of delimiter is specified,
apart from the fact that this can influence how many lines are processed (if any).

The error occurs only when Move fields is the filter type.

When one is copying a field [say] to the end of a line, there is no such error,
even when the line does not terminate with the specified delimiter.

Meanwhile, an obvious workaround to avoid this critical error is to add a right margin to the lines to be processed.

The right margin should just contain the specified field delimiter.

This can be readily removed after the lines have been processed.

Thanks David,

Fixed ready for v9.5 today or Monday.

Thanks. Just installed v9.5

David

Simon,

Thanks. The list index error issue has been fixed, but .....

Now that I've updated to version 9.5, I decided to test the same trial run filter.

On line 3, two fields are moved from the end, even though only one field should have been. Likewise, if you use comma as the delimiter, the same kind of bug occurs for the last line.

This is now a critical issue, because it will alter the behaviour of existing filters that move fields from the end.

David

END-1 .. END-0 actually specifies two fields, not one.

Having a trailing delimiter on a line is actually an empty field.

The question is more to do with backwards compatibility.

Please check whether before the out of bounds error was fixed, if moving END-1 ... END-0 used to move two fields or only one.

If the programming error also caused only one field to be moved before, then users need to be alerted.