Table of Contents

Name

tsd - scan and report contents of an HDF4 SDS

Synopsis

tsd [parameter=value]

Description

tsd lists the contents of a HDF Scientific Data Set (SDS). It can optionally print out the data (in ascii format) from all or specified SDS# to a specified file, with or without their coordinate values.

Parameters

The following parameters are recognized in any order if the keyword is also given:
in=hdf_file
Input file (HDF FLOAT32 SDS), no default.
out=table_file
Optional output file in which an ascii dump of all SDS, printed side by side, are dumped. Obviously in this mode all datasets must have the same rank and shape, selected SDS that are not the same size and rank as the first one, will not be displayed. Currently all the data are accumulated in memory. Default: none.
format=C_fmt
printf(3) -like Format specification, in which the data will be dumped if output requested. A space is automatically inserted between the columns representing an SDS. [Default: %g].
coord=t|f
Print out the coordinate system too? If selected, they will be printed in the first NDIM columns (except if dummy=f was selected, see below), followed by whichever columns were selected (see select= below). [Default: f].
select=
Select the SDS# by their position in the file, the first one being numbered 1. [Default: all].
dummy=t|f
Should coordinate values of dummy axes (of which the dimension is 1) be displayed? Default: t

Example


    % tsd hdf000bg
Found 3 scientific data sets in /d/hdf000bg
1: Y-VELOCITY AT TIME=0.00E+00(40,67) cm/sec  -> [2680 elements of type: 5
(FLOAT32)]
2: Z-VELOCITY AT TIME=0.00E+00(40,67) cm/sec  -> [2680 elements of type: 5
(FLOAT32)]
3: DENSITY AT TIME=0.00E+00(40,67) gms/cm**3  -> [2680 elements of type: 5
(FLOAT32)]
Note that the dimensions listed, (40,67) in the example above, are not the usual FORTRAN-convention. The last index (67) will be the one running fastest, i.e. data is stored (and dumped if requested) in column-major mode. For example, in the FITS conversion program (see sdsfits(1NEMO) ) the header is rearranged as to make it appear row-major, but the data is not.

Limitations

If multiple SDS exist, they must all have the same shape (or at least have as much data as the first SDS in the file), otherwise un-allocated memory is being accessed. Error control in that case depends on the operating system.

If output (ascii) requested, all data from all SDS is allocated in memory.

The datatype of the SDS must be FLOAT32 (normally in fortran: REAL*4, and in C: float)

The data is listed in the order of the file.

See Also

hdf(1) , hdfls(1) , fitssds(1NEMO) , sdsfits(1NEMO)

Author

Peter Teuben

Update History


25-dec-94    V1.0 Created    PJT
21-may-95    V1.1/2 fixed a bug, added new output features    PJT
27-aug-96    V1.3 header output now to stderr (dprintf)    pjt
10-dec-04    V1.5 added dummy=    PJT


Table of Contents