empty cells in CSV export

[ducin]

I'm developing a mechanism which could be exporting the same data either in xlsx or csv. My data structure is not a regular rectangle, unfortunately. If x is a single cell with content and _ is an empty cell, the whole sheet may look like:

xx__
xx__
xx__
____
xxxx
xxxx
xxxx

The exported excel is ok. But exported CSV creates blank cells which look like the following:

"Source:","my System","","","",""
"Data:","DataType","","","",""
"","","","","",""
01","13997506","90824763","6.488638976114745","143068242","10.22098093760417"

"2013- 02","13802513","89247193","6.466010428680632","139424347","10.10137407586575"

Is there any option to disable outputting blank cells in CSV? This doesn't seem to be difficult to code, but I couldn't find such things in the docs.

[MarkBaker]

We don't provide any method as standard to eliminate empty cells from the CSV output, or even just "trailing" empty cells. The former would provide an unpredictable csv file, readers wouldn't be able to identify columns cleanly, so probably isn't worthwhile; the latter is certainly possible

[ducin]

Thanks for your reply. So if I need such feature, what is the best approach to do this (from the architectural point of view)? I wouldn't like to parse the final csv output from your lib :) Is there a hook slot I can implement or something like that?

[MarkBaker]

Thanks for telling us that our CSV output is useless, sadly we provide a straightforward structure where, once you know how many entries are defined in one row, you know how many are defined in every row. I'd mistakenlky thought that this was the norm, because it's the behaviour reflected by most of the common spreadsheet programs.

The way to implement this would be to provide a new "trim" property in the CSV Writer, together with a getter and setter method for that property; then modify the -writeLine() method to trim trailing empty entries in the $pValues array, and the only to write the line if the count of $pValues entries is still greater than 0.

[ducin]

I didn't mean that your CSV output is useless (it was not an offence, to be honest). Probably most of CSV outputs need precise rectangular shape, but in my case, it's not, unfortunately.

In my case both excel and CSV consist of headers (meta-information about the document) and the real data (numbers). Meta-data is not who created the document and who modified it last time, but things like categories, period of time the data relates to (these can be different), structure indicators, etc.

In excel spreadsheet an empty cell is just an empty cell. But in CSV it is a little difference whether there is a ,"","", I hope you understand what I mean. Thanks you for instructions with the trim property!