write variables into seprate text file (essay style)
write variables into seprate text file (essay style)
Hi Blair,
what would be a great adddition to ML would be the possibility to write all data - ragardless of item type - into a seperate text file. As you know, we often use custom items (with HTML text area tags) to record large string files/essays. These are recorded only abbreviated (255?) into spss. If we want to code these variables separately we have to print out all the other data too or delete everything but the text passages. A separate file would come quite in handy.
Thanks,
Rene
Hmm. I can see why this would be useful, but I'm not sure how I would do it because ML doesn't know in advance what your custom item is doing or what kind of data you are expecting.
wouldn't it be possible to intercept the input type (text area)? ML could, then, identify it and write it separately...
No, because the posted command doesn't specify the type of data. By the time MediaLab gets it from the post command, everything is reduced to variable names and strings of text. Maybe an option like that for the custom item could be set using a paremeter value. Like the way you can specify a filename where fill-in-the-blank values can be saved.
But you can specify in ML if the posted from data is written as numeric or string. Why not add an additional option that lets you chose between string1 (within the raw data) and string2 (into an extra file) (and even better: string3 using both)?
Rene,
I think this might help:
http://www.empirisoft.com/medialab/help/data_faq.htm#data_vertical
As described, you can place the vertical.mdb file in your byVariableName data folder. It will store the value of every variable in a vertical Access database file. I just tried it with some very long fill-in-the-blank text responses and all were perfectly in tact. Haven't tried it with a custom item yet but I think it could serve as at least a short term solution for you.
Anyways see if it helps and let me know how it goes.
Blair,
I don't know why, but I haven't tried this workaround out yet. I'll see into it next week and report to you after that;)
All the best,
Rene
Will be curious to hear if it's helpful.
