opts
settings.Read the data accoring to the settings opts$READ.*
and opts$TST.COL
(see Details).
tdmReadDataset(opts)
list of options, we need here
READ.TrnFn
: [tdmReadTrain] function with argument opts
for reading
the training data and returning them in a data frame
READ.NROW
: [-1] read only that many rows from opts$filename. -1 for 'read all rows'.
READ.TstFn
: [NULL] function with argument opts
for reading
the test data and returning them in a data frame. If NULL
then skip test file reading.
TST.COL
: ["TST.COL"] string, create a column with the name of this string
in dset
, which has 0 for training and 1 for test data
path
: used by READ.TrnFn to locate file
dir.data
: used by READ.TrnFn to locate file
dset
, a data frame with all data read
When opts$READ.TstFn==NULL
, then only opts$READ.TrnFn
is used.
When opts$READ.TstFn!=NULL
, the following things happen in tdmReadDataset
:
Data are read from opts$filename
and from opts$filetest
. Both data sets are bound
together, with a new column opts$TST.COL
having '0' for the data from opts$filename and
having '1' for the data from opts$filetest. The branch using opts$TST.COL is invoked either with
umode="TST" in unbiasedRun
or with opts$TST.kind="col" in tdmModCreateCVindex
.