Problems converting data from CSV to SPSS

Hi everyone, I hope you’re doing great.
We developed a survey with more than 700 variables in spanish, everything was OK until we have to convert data into Spss.
I’m downloading the data in CSV format and then imported it in SPSS, but in some cases the data haved mix and do not share the format from previous cases like in the picture.

Could you please help me fix this out please?
Thank you

Welcome to the community, @martin_maranon! Have you tried with the XLS (legacy) format?

Thank you @Kal_Lam, congratulations for all the work you’ve bee doing here.
I will, it is better to work with XLS (legacy) or with XLS format?

Thanks again.

1 Like

Hello,
Our experience is to work best with xls (xml header and values) export, plus SPSS Labels export. You don’t need to go with csv or legacy.

Some further hints in a recent posting: When you create question with type multiple selection not support for spss syntax file - #8 by wroos.
Would like to keep in touch for further exchange on SPSS transfer…

@Kal_Lam? Any reason why to use legacy export? (I would also suggest to change it in the KoBo support article: Converting Data into SPSS and/or Stata — KoBoToolbox documentation (Which is talking of CSV even).

More discussion in the forum can be found with search for “SPSS”.

2 Likes

Cool, thank you @wroos!
For the type of variables we are managing, the statistics we need to make and the time we have left, I will love to work with the XLS (labels) format.
Wonderful results with categorical variables, but I have some troubles with the numeric variables, the variable’s values appears in Excel in the numeric format, but when I want to make some calculations Excel doesn’t recognise the values, and is not aviable to make statistics, like the mean.


What dou you say?
any experience with this format?
Thank you very much.

Thanks @Kal_Lam, the other formats in XLS shows the data correctly.
It seems that the CSV format have some trouble with some spanish characters, like “ñ” or the “accent mark” (á,ó,é,í,ú…), that doesn’t appear in english lenguage.

CSV file
image
XLS formats
image
image

Do you believe this can be fixed by changing the lenguage settings of the project? or this is a regular issue?
Can you please help me with the differences between XLS and XLS(legacy) format?

Thank you verymuch again!

1 Like

Hello,
I enhanced the posting above.
Just go with normal XLS export (NOT legacy), I would suggest. We are using it extensively, incl. Arabic texts.

For real numeric variables in form design, unfortunately exported as text by KoBo, we use Excel conversion “Text to columns” (Data menu), selecting the column. (Partly also for date values.). Simple format change seems not enough, will only change cell alignment.

There is also a KoBo enhancement request:

Kind regards

1 Like

Thank you,
I will love to keep in touch in any issues related with gathering and processing the data…

I agree, taking a look to all the options mentioned, XLS (XLM header and values) export seems the best way to have manageable data, at least in my case (we have several questions groups, and different types of data. The legacy formart seems to complicate the variable names).

Validating the data in the platform does modify in any way the data retrived (format, including a nex variable…)?

I’ll let you know if I have any issues with the spss syntax for labeling and formating the variables, hope not!

Thanks again, a pleasure to talk to you.
Cheers!

2 Likes

@wroos, I just wanted to see if the legacy export too had any issues?

1 Like

The XLS format is the download supported by the KPI while the XLS (Legacy) is a format supported by KC.

Hi @martin_maranon @wroos
We have been having an issue with the CSV export not only for the Spanish language but also English. The line breaks introduced due to the carriage return key Text data type and csv data export - #14 by cp622

In the meantime, what @wroos recomended has actually proven very useful even on my end. I admit we have taken sometime to update the write up on the support around the use of CSV and the XLS downloads. This is largely because CSV woudl still be the best option.

In terms of action points, we will be looking at updating the post and toying around with various options of download to fully utilize the conversion to SPSS capabilities.

Stephane

1 Like

Hello @stephanealoo
Why do you think CSV transfer to Excel is “still the best option” for transfer to SPSS?

See XLS > SPSS:
https://www.ibm.com/docs/en/spss-statistics/27.0.0?topic=files-reading-excel
and for CSV there:
“Note: The Import Data function does not support CSV data that includes embedded line breaks in quoted text. A possible workaround is to save the CSV file (that includes quoted embedded line breaks) as an .xls/.xlsx file, and then use the import Excel feature.”

Hi Wroos,
Its all about data management and upload interfaces which CSV has a better versatility compared to XLS provides. Preference of CSV and XLS is however a much more subjective choice that from a user support perspective, I would rather defer users to their organizational needs.

Stephane

Hello Stephane,
Ok, in general. But for SPSS XLSX seems the better and more robust choice, as far as we experienced…