The DCX$ANALYZE_INIT routine initializes the context area for a
statistical analysis of the data records to be compressed.
Format
DCX$ANALYZE_INIT context [,item_code ,item_value]
1 – Returns
OpenVMS usage:cond_value
type: longword (unsigned)
access: write only
mechanism: by value
Longword condition value. Most utility routines return a
condition value in R0. Condition values that this routine can
return are listed under Condition Values Returned.
2 – Arguments
context
OpenVMS usage:context
type: longword (unsigned)
access: write only
mechanism: by reference
Value identifying the data stream that DCX$ANALYZE_INIT
initializes. The context argument is the address of a longword
containing this value. DCX$ANALYZE_INIT writes this context
into the context argument; you should not modify its value. You
can define multiple context arguments to identify multiple data
streams that are processed simultaneously.
item_code
OpenVMS usage:longword_unsigned
type: longword (unsigned)
access: read only
mechanism: by reference
Item code specifying information that you want DCX$ANALYZE_INIT
to use in its analysis of data records and in its computation
of the mapping function. DCX$ANALYZE_INIT reads this item_code
argument, which is the address of the longword contained in the
item code.
For each item_code argument specified in the call, you must
also specify a corresponding item_value argument. The item_value
argument contains the interpretation of the item_code argument.
The following symbolic names are the five legal values of the
item_code argument:
DCX$C_BOUNDED
DCX$C_EST_BYTES
DCX$C_EST_RECORDS
DCX$C_LIST
DCX$C_ONE_PASS
item_value
OpenVMS usage:longword_unsigned
type: longword (unsigned)
access: read only
mechanism: by reference
Value of the corresponding item_code argument. DCX$ANALYZE_INIT
reads the item_value argument, which is the address of a longword
containing the item value.
The item_code and item_value arguments always occur as a pair,
and together they specify one piece of "advice" for the DCX
routines to use in computing the map function. Note that, unless
stated otherwise in the list of item codes and item values, no
piece of "advice" is binding on DCX; that is, DCX is free to
follow or not to follow the "advice."
The following table shows, for each item_code argument, the
possible values for the corresponding item_value argument:
Item Code Corresponding Item Value
DCX$C_BOUNDED A Boolean variable. If bit <0> is true (equals
1), you are stating your intention to submit
for analysis all data records that will be
compressed; doing so often enables DCX to
compute a better compression algorithm. If
bit <0> is false (equals 0) or if the DCX$C_
BOUNDED item code is not specified, DCX computes
a compression algorithm without regard for
whether all records to be compressed will also
be submitted for analysis.
DCX$C_EST_BYTES A longword value containing your estimate of
the total number of data bytes that will be
submitted for compression. This estimate is
useful in those cases where fewer than the total
number of bytes are presented for analysis. If
you do not specify the DCX$C_EST_BYTES item
code, DCX submits for compression the same
number of bytes that was presented for analysis.
Note that you may specify DCX$C_EST_RECORDS or
DCX$C_EST_BYTES, or both.
DCX$C_EST_ A longword value containing your estimate of
RECORDS the total number of data records that will be
submitted for compression. This estimate is
useful in those cases where fewer than the total
number of records are presented for analysis.
If you do not specify the DCX$C_EST_RECORDS
item code, DCX submits for compression the same
number of bytes that was presented for analysis.
DCX$C_LIST Address of an array of 2*n+1 longwords. The
first longword in the array contains the value
2*n+1. The remaining longwords are paired; there
are n pairs. The first member of the pair is
an item code, and the second member of the pair
is the address of its corresponding item value.
The DCX$C_LIST item code allows you to construct
an array of item-code and item-value pairs and
then to pass the entire array to DCX$ANALYZE_
INIT. This is useful when your language has
difficulty interpreting variable-length argument
lists. Note that the DCX$C_LIST item code may be
specified, in a single call, alone or together
with any of the other item-code and item-value
pairs.
DCX$C_ONE_PASS A Boolean variable. If bit <0> is true (equals
1), you make a binding request that DCX make
only one pass over the data to be analyzed. If
bit <0> is false (equals 0) or if the DCX$C_
ONE_PASS item code is not specified, DCX may
make multiple passes over the data, as required.
Typically, DCX makes one pass.
3 – Description
The DCX$ANALYZE_INIT routine initializes the context area for a
statistical analysis of the data records to be compressed. The
first (and typically the only) argument passed to DCX$ANALYZE_
INIT is an integer variable to contain the context value. The DCX
facility assigns a value to the context variable and associates
the value with the created work area. Each time you want a
record analyzed in that area, specify the associated context
variable. You can analyze two or more files at once by creating a
different work area for each file, giving each area a different
context variable, and analyzing the records of each file in the
appropriate work area.
4 – Condition Values Returned
DCX$_INVITEM Error; invalid item code. The number of
arguments specified in the call was incorrect
(this number should be odd), or an unknown
item code was specified.
DCX$_NORMAL Normal successful completion.
This routine also returns any condition values returned by
LIB$GET_VM.