Names of Matrix Variables from MGET (MATRIX-END MATRIX command)

• The MGET statement automatically creates matrix variable names for the matrices it reads.
• All new variables created by MGET are reported to the user.
• If a matrix variable already exists with the same name that MGET chose for a new variable, the new variable is not created and a warning is issued. The RELEASE statement can be used to get rid of a variable. A COMPUTE statement followed by RELEASE can be used to change the name of an existing matrix variable.

MGET constructs variable names in the following manner:

• The first two characters of the name identify the row type. If there are no cells and no split file groups, these two characters constitute the name:

  CV . A covariance matrix (rowtype COV)

  CR . A correlation matrix (rowtype CORR)

  MN . A vector of means (rowtype MEAN)

  SD . A vector of standard deviations (rowtype STDDEV)

  NC . A vector of numbers of cases (rowtype N)

  CN . A vector of counts (rowtype COUNT)

• Characters 3–5 of the variable name identify the cell number or the split-group number. Cell identifiers consist of the letter F and a two-digit cell number. Split-group identifiers consist of the letter S and a two-digit split-group number; for example, MNF12 or SDS22.
• If there are both cells and split groups, characters 3–5 identify the cell and characters 6–8 identify the split group. The same convention for cell or split-file numbers is used; for example, CRF12S21.
• After the name is constructed as described above, any leading zeros are removed from the cell number and the split-group number; for example, CNF2S99 or CVF2S1.