MSAVE Statement: Writing Matrix Data Files (MATRIX-END MATRIX command)
The MSAVE
statement writes matrix expressions to a matrix-format data file
that can be used as matrix input to other procedures. (See MATRIX DATA for a discussion of matrix-format
data files.) The syntax of MSAVE
is as follows:
MSAVE matrix expression
/TYPE = {COV }
{CORR }
{MEAN }
{STDDEV}
{N }
{COUNT }
[/OUTFILE = 'filespec']
[/VARIABLES = variable list]
[/SNAMES = variable list]
[/SPLIT = split vector]
[/FNAMES = variable list]
[/FACTOR = factor vector]
- Only one matrix-format data file can be saved in a single matrix program.
- Each
MSAVE
statement writes records of a single rowtype. Therefore, severalMSAVE
statements will normally be required to write a complete matrix-format data file. - Most specifications are retained from one
MSAVE
statement to the next so that it is not necessary to repeat the same specifications on a series ofMSAVE
statements. The exception is theFACTOR
specification, as noted below.
Example
MSAVE M /TYPE=MEAN /OUTFILE=CORRMAT /VARIABLES=V1 TO V8.
MSAVE S /TYPE STDDEV.
MSAVE MAKE(1,8,24) /TYPE N.
MSAVE C /TYPE CORR.
- The series of
MSAVE
statements save the matrix variables M, S, and C, which contain, respectively, vectors of means and standard deviations and a matrix of correlation coefficients. The matrix-format data file thus created is suitable for use in a procedure such asFACTOR
. - The first
MSAVE
statement saves M as a vector of means. This statement specifiesOUTFILE
, a previously defined file handle, andVARIABLES
, a list of variable names to be used in the IBM® SPSS® Statistics data file. - The second
MSAVE
statement saves S as a vector of standard deviations. Note that theOUTFILE
andVARIABLES
specifications do not have to be repeated. - The third
MSAVE
statement saves a vector of case counts. The matrix functionMAKE
constructs an eight-element vector with values equal to the case count (24 in this example). - The last
MSAVE
statement saves C, an 8 × 8 matrix, as the correlation matrix.