|
<< Click to Display Table of Contents >> Navigation: Reference Manual > Files > CSV type data file |
CSV and XLSX files are always saved as CSV Portable files where the headings are the entry names and the serial number is put in the first column with "Serial" as the heading:
•Single-coded questions are stored as a response number
•Multi-coded questions are stored as response numbers separated by spaces
•Integer questions are stored as numbers, possibly with a leading - (minus), or U (undefined)
•Float questions are stored as numbers, possibly with decimal places and a leading - (minus), or U (undefined)
•Character and verbatim questions are stored as text
Data export can be used to produce other methods of storing the data in a data file.
When reading CSV or XLSX data files other options are possible for single-coded and multi-coded questions:
For single-coded questions the program will accept:
•a number which will be treated a the response number.
•text answers where the text matches the text of a response, not case sensitive.
For multi-coded questions with only one column the program will only accept:
•The number "0" (zero). This will not be used.
•Numbers separated by spaces which will be treated as response numbers.
Many columns can be used for a multi-coded question. The columns must all be headed by the question name followed by a special character (usually underscore) followed by a number.
For example, question Q5 can be stored in columns headed Q5_1, Q5_2, Q5_3, and so on.
With heading like this the program will accept in any of the columns:
•The number "0" (zero). This will not be used.
•The number "1" (one). This will be used as the response number added to the header for this column.
•A number greater than 1. It does not matter in which column this number is in, it will be used as a response number.
•Text answers where the text matches the text of a response, not case sensitive. It does not matter which column this number is in, it will be used as a response number with the matching text.
Not all responses must have a header, for example Q3 could have 9 responses and you could have Q3_1, Q3_2, and Q3_9. A "1" in the Q3_9 column will be treated as response 9.
Because the program will accept response text as valid data when reading files there can be a problem if a response text is a number (without any additional text) that is not the response number for the response.
For example, if response number 5 has a response text of "6" then the program will not know whether a "6" in the data is response 6 or response 5. It will assume the number is the response number if it is in the valid range of response numbers.
The program will warn about ambiguous number texts unless "Suppress ambiguous response text warnings" is set in Project global settings.