Hi,

I can not get the 'Feeds Excel' parser working. I installed and enabled the module, downloaded the library as the README suggests, and set up an importer as usual. So far, everything is working as expected; however, I don't understand how the mapping is supposed to work for an Excel sheet.

I don't get how the provided tokens are to be used, and if I enter "cell addresses" in the mapping source field, the module seems not to iterate through the proper sheet cells/rows/colums (e.g. if I enter the address C, or C3, or C:3 for "title", I get 14 nodes with the title "C", respectively "C3", respectively "C:3", if the sheet has 14 rows).

Example: A sheet has an image in the first row and column titles in the second row, so I want the import to ignore those two rows and to start in the third row (Excel Parser - Settings: "Mapping mode: Rows"; "Fixed Range: 3:1000"; "Iterative range: 3:1000"). The sheet has a title in column "C", a date in column "I" and some text in column "N", but there's no token for column "C", and addresses like C, [C], C3, [C3], C:3, or [C:3] as mapping sources don't work. So how do I process a sheet like this?

Thanks, -asb

CommentFileSizeAuthor
#3 strange-characters-after-import.jpeg31.29 KBasb

Comments

asb’s picture

asb commented
Component: User interface » Documentation

I did some more guesswork; it seems that the spreadsheet's column names "A", "B", "C", etc. simply are not supported, and the provided tokens are just examples. In my case, something like this seems to work:

[column-3] -> Title
[column-3] -> Body
[column-14-raw] -> Taxonomy: Foo
[column-19-raw] -> Taxonomy: Bar

For text, the "raw" token values shouldn't be used as this seems to insert one "special character" between every "real" character.

Edit: Wrong observation, no matter if I use the raw or the "normal" token, I'm sometimes getting bogus characters in the node title, in CCK text field, or the node body. Example:

�u�m� �z�u� �ü�b�e�r�l�e�b�e�n�!� � �

Whatever these characters are, they make the module pretty useless (screenshot attached). To make this even stranger, I'm getting these borked strings unsystematically; some rowas from the spreadsheet are just fine. Maybe it is necessary to somehow clean up the encoding when reading the spreadsheet?

Also, for whatever reason, 'Feeds Tamper' seems not to work together with the 'Excel Parser' module, so I can not clean up messy spreadsheets on the fly (data in spreadshees is *alway* messy ;)

It'd be nice if things like this could be documented somewhere.

derhasi’s picture

derhasi commented

For text, the "raw" token values shouldn't be used as this seems to insert one "special character" between every "real" character. For whatever reason, "Feeds Tamper" seems not to work together with the "Excel Parser", so I can not clean up messy spreadsheets on the fly (data in spreadshees is *alway* messy ;)

This might have to do somethind with the excel file encoding. It's pretty hard to solve this with the "old" library 1.x is using.
What "special character" is it?

asb’s picture

asb commented
StatusFileSize
new strange-characters-after-import.jpeg31.29 KB

Screenshot of how the strings look after an import.

mrromios’s picture

mrromios commented
Issue summary: View changes
Status: Active » Closed (won't fix)