The CDF functions/procedures available at the
IDL command line exactly mirror the
Standard and Internal Interface functions available
for a program written in C.
The CDF C Reference Manual
describes the functionality of each in detail. The
syntax required when calling each function is as follows...
IDL> status = CDFlib (op1, ...)
IDL> status = CDFcreate (CDFpath, numDims, dimSizes, encoding, majority, $
IDL> id)
IDL> status = CDFopen (CDFpath, id)
IDL> status = CDFdoc (id, version, release, copyright)
IDL> status = CDFinquire (id, numDims, dimSizes, encoding, majority, $
IDL> maxRec, numVars, numAttrs)
IDL> status = CDFclose (id)
IDL> status = CDFdelete (id)
IDL> status = CDFerror (status, text)
IDL> status = CDFattrCreate (id, attrName, attrScope, attrNum)
IDL> status = CDFattrRename (id, attrNum, attrName)
IDL> status = CDFattrInquire (id, attrNum, attrName, attrScope, maxEntry)
IDL> status = CDFattrEntryInquire (id, attrNum, entryNum, dataType, $
IDL> numElems)
IDL> status = CDFattrPut (id, attrNum, entryNum, dataType, numElems, value)
IDL> status = CDFattrGet (id, attrNum, entryNum, value)
IDL> status = CDFvarCreate (id, varName, dataType, numElems, recVary, $
IDL> dimVarys, varNum)
IDL> status = CDFvarRename (id, varNum, varName)
IDL> status = CDFvarInquire (id, varNum, varName, dataType, numElems, $
IDL> recVary, dimVarys)
IDL> status = CDFvarPut (id, varNum, recNum, indices, value)
IDL> status = CDFvarGet (id, varNum, recNum, indices, value)
IDL> status = CDFvarHyperPut (id, varNum, recStart, recCount, recInterval, $
IDL> indices, counts, intervals, buffer)
IDL> status = CDFvarHyperGet (id, varNum, recStart, recCount, recInterval, $
IDL> indices, counts, intervals, buffer)
IDL> status = CDFvarClose (id, varNum)
IDL> attrNum = CDFattrNum (id, attrName)
IDL> varNum = CDFvarNum (id, varName)
IDL> status = CDFlib (op1, ...)
IDL> epoch = computeEPOCH (year, month, day, hour, minute, second, $
IDL> millisecond)
IDL> EPOCHbreakdown, epoch, year, month, day, hour, minute, second, $
IDL> millisecond
IDL> epoch = parseEPOCH (epochString)
IDL> encodeEPOCH, epoch, epochString
IDL> encodeEPOCH1, epoch, epochString
IDL> encodeEPOCH2, epoch, epochString
IDL> encodeEPOCH3, epoch, epochString
The IDL data types to use for integer arguments are always longwords or
arrays of longwords using IDL's data type
of TYP_LONG. Attribute entries and variable values should have the
data types corresponding their CDF data types. CDF pathnames and
variable/attribute names are IDL character strings.
CDFid and CDFstatus are also longwords. The CDF C Reference Manual
may be used as a guide (especially for CDFlib). Arguments passed
to the CDF functions/procedures must
obviously be created and initialized before the call. Arrays of longwords are
necessary for dimension sizes, variances, etc. Arguments
returned by
the CDF functions will be created to be of the proper data type and size
(they do not have to be created before calling the CDF function).
NOTE: There is one difference between how the Internal Interface (CDFlib) is used in IDL and in a C/Fortran application. In IDL the current CDF must be selected each time CDFlib is called. Failure to select the current CDF may result in IDL crashing.
Also provided are two functions, row_to_col and col_to_row, that may be used to change the majority of an array (or arrays) of values. This function is necessary because CDF can store variable values in either row or column major order. The syntax of row_to_col is as follows...
IDL> status = row_to_col (inArray, outArray, numDims, dimSizes, numBytes, $
arrayCount
where...
The array (or arrays) of values to convert --- row major ordering. Any valid IDL data type is allowed.
The converted array (or arrays) of values --- column major ordering. This array will be created by row_to_col.
The number of dimensions in the array (or arrays). This value must be a longword.
The size of each dimension. This must be an array of longwords --- one longword per dimension.
The number of bytes in each value. This value must be a longword.
The number of arrays to convert. The order of the arrays will not be affected. This value must be a longword.
Completion status code. This value will be a longword. The possible values are defined in cdf1.pro.
Multiple arrays (records) of CDF variable values should always be stored with the last dimension of the IDL array corresponding to the record number. For example, assume a CDF variable with two dimensions whose sizes are [180,360]. If 50 records are to be stored, an IDL array with three dimensions whose sizes are [180,360,50] should be used. If the CDF variables have row major ordering, row_to_col would be called with numDims set to 2, dimSizes set to [180,360], and arrayCount set to 50.
The syntax of col_to_row is as follows...
IDL> status = col_to_row (inArray, outArray, numDims, dimSizes, numBytes, $
arrayCount
where...
The array (or arrays) of values to convert --- column major ordering. Any valid IDL data type is allowed.
The converted array (or arrays) of values --- row major ordering. This array will be created by col_to_row.
The number of dimensions in the array (or arrays). This value must be a longword.
The size of each dimension. This must be an array of longwords --- one longword per dimension.
The number of bytes in each value. This value must be a longword.
The number of arrays to convert. The order of the arrays will not be affected. This value must be a longword.
Completion status code. This value will be a longword. The possible values are defined in cdf1.pro.
col_to_row would be used in those cases where an array of values created by IDL (in column-major ordering) are to be written to a row-major CDF.