writer#

Purpose#

Writes a matrix to a GAUSS dataset.

Format#

ret = writer(fh, x)#

Parameters:

fh (scalar) – handle of the file that data is to be written to
x (NxK matrix) – data

Returns:

ret (scalar) – specifying the number of rows of data actually written to the dataset.

Examples#

Create and write to dataset#

x = { 1 2 3,
      4 5 6 };

var_names = "alpha" $| "beta" $| "gamma";
nvars = rows(var_names);

// Create dataset and get file handle
fh = dataCreate("test.dat", var_names, nvars, 8, 1);

// Write data in 'x' to test.dat
ret = writer(fh, x);

close(fh);

Create and iteratively write to dataset#

new;

nvars = 10;

// Create dataset and get file handle
fp = dataCreate("data.dat", "X", nvars, 8, 1);

// Check to confirm that data file was created
if fp == -1;
   errorlog "Can't create output file";
   end;
endif;

for i(1, 10, 1);
   y = rndn(100,10);
   ret = writer(fp,y);

   // Check to see if all data was written
   if ret != rows(y);
     errorlog "Write failed";
     fp = close(fp);
     end;
   endif;

endfor;

fp = close(fp);

In this example, a 1000x10 dataset of Normal random numbers is written to a dataset called data.dat. The variable names are X01 - X10.

Remarks#

saved() provides a more convenient method to create a dataset.
The file must have been opened with dataCreate(), or dataOpen() for update or append. The keywords create and open are still supported as well.
The data in x will be written to the dataset whose handle is fh starting at the current pointer position in the file. The pointer position in the file will be updated, so the next call to writer() will put the next block of data after the first block. (See open and create for the initial pointer positions in the file for reading and writing.)
x must have the same number of columns as the dataset. colsf() returns the number of columns in a dataset.
If the dataset is not double precision, the data will be rounded as it is written out.
If the data contain character elements, the file must be double precision or the character information will be lost.
If the file being written to is the 2-byte integer data type, then missing values will be written out as -32768. These will not automatically be converted to missings on input. They can be converted with the miss() function:
```
x = miss(x,-32768);
```
Trying to write complex data to a dataset that was originally created to store real data will cause a program to abort with an error message. (See create for details on creating a complex dataset.)