This is the personal project of Tom de Neef. This project is created automatically during user registration as an easy place to store personal data that doesn't need its own project such as cloned repositories.
Hi,
I am not sure if you are still interested in the SPSSio unit that you have released. But anyway. I have been trying to use it, and in general it works fine.
However, it crashes with some files that SPSS itself is able to read. The SPSS DLLs (I have a Delphi wrapper for them) are also able to read the files. Foreign in R also reads the files.
The problem comes in two flavors: one is that TSPSSio.PrepareReading detects unkown rectypes (much as the Foreign and Haven packages in R often do). If I change the code to ignore such record types, rather than throw an error, then everything is fine.
Another is like the other, but PrepareReading also miscounts the number of variables. It is possible to get it to read the right number of variables, but the data are then gibberish. Howevr, Foreign reads the file just fine.
Opening and saving the file in SPSS (v27) helps somewhat . The number of variables becomes correct, but data are still gibberish.
In case you have any interest still, I enclose a file. (It is a birth history file from UNICEFs MICS surveys - the program I am using it with is a variance estimation program for such data).
All the best,
Jon Pedersen
Hi Jon,
good to hear that my past work may have been of use.
I developed the tSPSSio units 10 years ago. I am now 78. It has been too
long ago to dive into it again.
Makes me wonder if I should remove my contributions from SourceForge.
Anyway, sorry that I can't help.
Cheers
Tom
Hi,
I am not sure if you are still interested in the SPSSio unit that you have
released. But anyway. I have been trying to use it, and in general it works
fine.
However, it crashes with some files that SPSS itself is able to read. The
SPSS DLLs (I have a Delphi wrapper for them) are also able to read the
files. Foreign in R also reads the files.
The problem comes in two flavors: one is that TSPSSio.PrepareReading
detects unkown rectypes (much as the Foreign and Haven packages in R often
do). If I change the code to ignore such record types, rather than throw an
error, then everything is fine.
Another is like the other, but PrepareReading also miscounts the number of
variables. It is possible to get it to read the right number of variables,
but the data are then gibberish. Howevr, Foreign reads the file just fine.
Opening and saving the file in SPSS (v27) helps somewhat . The number of
variables becomes correct, but data are still gibberish.
In case you have any interest still, I enclose a file. (It is a birth
history file from UNICEFs MICS surveys - the program I am using it with is
a variance estimation program for such data).
All the best,
Jon Pedersen
Hi,
I am not sure if you are still interested in the SPSSio unit that you have released. But anyway. I have been trying to use it, and in general it works fine.
However, it crashes with some files that SPSS itself is able to read. The SPSS DLLs (I have a Delphi wrapper for them) are also able to read the files. Foreign in R also reads the files.
The problem comes in two flavors: one is that TSPSSio.PrepareReading detects unkown rectypes (much as the Foreign and Haven packages in R often do). If I change the code to ignore such record types, rather than throw an error, then everything is fine.
Another is like the other, but PrepareReading also miscounts the number of variables. It is possible to get it to read the right number of variables, but the data are then gibberish. Howevr, Foreign reads the file just fine.
Opening and saving the file in SPSS (v27) helps somewhat . The number of variables becomes correct, but data are still gibberish.
In case you have any interest still, I enclose a file. (It is a birth history file from UNICEFs MICS surveys - the program I am using it with is a variance estimation program for such data).
All the best,
Jon Pedersen
Hi Jon,
good to hear that my past work may have been of use.
I developed the tSPSSio units 10 years ago. I am now 78. It has been too
long ago to dive into it again.
Makes me wonder if I should remove my contributions from SourceForge.
Anyway, sorry that I can't help.
Cheers
Tom
Op wo 15 mei 2024 10:23 schreef Jon Pedersen jonp85@users.sourceforge.net: