davidearl / jcomma Goto Github PK

(a) add the ability to determine header rows based on content (either or both of the last header line and the first content line)
(b) some flexibility as to which header line(s) constitute the actual column headers

When a field is composed from more than one column or field, the append comma and append space options are being applied to the entire string so far, so the effect is cumulative, rather than being applied to each input individually.

To help those who might wish to contribute, it would be helpful to have some kind of testing mechanism - either unit tests or even a set of regression tests.

They would provide useful examples as a kind of documentation as well.

e.g. replace regexp with value blank

As it says...

User request: allow recipe JSON to be pretty printed

To avoid a lot of scrolling, keep the DO IT box at the top of the window at all times

They are in circles, which is fine, but the characters are small and not central

For example, if the same column in the previous row is equal to the current row, omit record.

Motivated by CAF statements and similar which have inter-alia the following structure:

Gross amount, Fee, Net amount, Transfer ID, Transfer amount

It might be useful to offer a SQL INSERT statement output format.

When the value for a condition comes from a column, it should have a qualifying row number for when multiple rows are being consumed. Possibly only display if overall rowCount is > 1 (ditto selecting from column).

... In local Storage by name, so they can be selected from a pull down

Would be useful to be able to sort the final output based on some part of the output

IE 11.589.10586.0 - Windows 10.

Untick "include header row" and reselect saved recipe, and box is ticked again

The column label 'AA' (and I think BA, CA etc), and any headings that refer to those columns (27, 53, ...) says 'invalid column letter...'

It looks like it's because I have a strpos(...) == FALSE rather than strpos(...) === FALSE for the second letter if there is one.

When a field is composed of more than one column or field, it would be useful to be able to prefix a minus sign to the input (when not blank), in the same way as commas and spaces can be appended currently. This would provide an easier way to deal with separate Credit and Debit columns in the input, when one or the other is always blank. A trim option would deal with the case when the "blank" column actually has some spaces. This doesn't deal with the case where the ideally blank input is actually set to zero; that will have to be dealt with as two output fields of the same name, where one or other is omitted on output.

GeoffG said by email:

When entering a regular expression in Jcomma, such as "replace using regular
expression", the box is rather small and given the nature of regexps one is
faced with a window into what may look like gibberish to the un-initiated!
Trying to review a regexp which doesn't quite do what you want can be a bit
tricky.

Would it be possible to make the regexp box(es) bigger? I could imagine having a
whole line for it, although it would be a bit trickier to make the layout look
nice.

The rest of the spec updates but not the json in the copy textarea