WriteCellRef only if _needsRef is set to true about largexlsx HOT 11 CLOSED

@salvois Related to this issue, I found that the xlsx file generated by LargeXlsx works perfectly in Excel, but failed to read using a certain go-lang library (excelize) because it needs every <row> to have r attribute. So I think BeginRow also need an optional parameter to always write the r attribute for that row.

from largexlsx.

@salvois , sorry for the delay.

I agree, an XLSX file with no cell reference is a perfectly valid XLSX file.

I had a look at how do we use Open XML SDK to read an XLSX file and why we need cell reference.

OPEN XML SDK itself do not need cell reference and can read a XLSX file without this information.

Our algorithm which use OPEN XML needs it. Maybe i could rewrite it without using cell reference but i'm not sure it is possible and i won't spend time on it for now.

Anyway, as you said, the is optional and i think LargeXlsx library should be able to optionaly write it because it should not decide what the result should be.
In other words, if the LargeXLSX user wants the attribute, for whatever reason, and because it is a valid argument, the library should write it. I don't think the library should decide what the result should be.

But you're the LargeXLSX owner and maybe do not agree with my opinion and because i've found a workaround, it is not a critical subject.

Thanks,

Mikk

from largexlsx.

Hi @Mikk182 ,
I'm sorry but I'm not sure I understand your proposal about the constructor accepting a needRef flag.
Could you elaborate and maybe provide an example on how you would use that, and why you would need this, please?
Thanks,
Salvo

from largexlsx.

Hi @salvois,

I would like to create a new branch and make a PR but seems I do not have right. So i've attached a zip file of the solution with changes.

I need writing cell reference because some of my generated files may be readen by DocumentFormat.OpenXml library and need this information.

Thanks

Mikk

LargeXlsx.zip

from largexlsx.

I finally create a Pull Request

from largexlsx.

Thanks for your pull request @Mikk182 .
At the moment, it's still not clear to me why you would need this, though.
To my knowledge, an XLSX file with no cell reference is a perfectly valid XLSX file, thus I'm very curious to know how the Open XML SDK is requiring it and what kind of problem you are running into with LargeXlsx as is. Very likely I'm missing some piece of information.
Thanks,
Salvo

from largexlsx.

@abrari-ipb thanks for this information. In my understanding, ISO/IEC 29500-1:2016 specifies that the r attribute of row and cell elements is optional, thus I'd dare to say that a library requiring it is not standard compliant and maybe this should be reported as an issue to that library.
I appreciate that workarounds may be useful sometimes, though, so I'll think about it, but I'm not very fond of "quirks modes" generally.

from largexlsx.

Hi @Mikk182 , did you have a chance to have a look at:

To my knowledge, an XLSX file with no cell reference is a perfectly valid XLSX file, thus I'm very curious to know how the Open XML SDK is requiring it and what kind of problem you are running into with LargeXlsx as is.

Thanks,
Salvo

from largexlsx.

Actually this is what is causing me issues with SAP: #17 (comment)

from largexlsx.

Hello @Mikk182 , thanks for your response, no need to be sorry.
I see your point now. While I appreciate you have a way to read XLSX files that work for you, you are accepting that you are requesting files that have stricter requirements than the standard, thus you may as well incur in other writers that do omit row and cell references.
It is still not very clear to me how "de facto standard" this stricter requirement is: as I said, I'm afraid of including compatibility or quirks modes for specific case, because they could easily proliferate. I'm not ruling that out though, just collecting information at the moment.
Concerning @natalie-o-perret issue, this puzzles me because her issue predates the release which included the optimization to omit row and cell references, so I'm waiting for more information about that too.
Thanks everyone for the moment,
Salvo

from largexlsx.

Hello @Mikk182, I have merged your pull request and adapted it to specify the new setting while constructing the XlsxWriter rather than per worksheet.
That feature will be available in the next release on LargeXlsx.
Please feel free to have a look, so that I can close this as resolved.
Thanks,
Salvo

from largexlsx.