Survo data files consist of header records which describe the
structure of the file and of any number of records for observations.
The data sets are saved in these files in the form of a data matrix
having, say, M columns (M=# of variables) and N rows (N=# of observat-
ions).
The variables can be either numeric or string variables.
A numeric variable can be any of 4 different types:
length
in bytes:
1: 0<=X<=254 (integer) missing value=255
2: -32768<=X<=32766 (integer) missing value=32767
4: real, 7 digits missing value=1e+38
8: real, 15 digits missing value=1e+306
A string variable can have a length varying from 1 to 70.
Each observation consisting of M variables (fields) needs a fixed
number of bytes. This number as well as the maximum number of fields
(variables) in each observation (record) is given by the user when
creating a data file (see FILE CREATE).
Since the values of the numeric variables are compressed into a binary
form in saving, a Survo data file cannot be listed as a text file.
To see its contents and to edit it, use FILE EDIT, for example.
The order of variables and variable types is immaterial. However, it
is a convention to define the first field as a string variable and
store the name of the observation in it.
The names of the variables (fields) are strings whose lengths vary
between 8 and 64. The 8 first characters are used as a reference
word in statistical operations. The rest of the name can be specified
freely, but it may include (in case of a numeric variable)
- the format of the variable for printouts etc. in the form (###.#)
- the range in the form {lower limit,upper limit}
1 = Creating a Survo data file (FILE CREATE)
Help System (web edition)