Subject: ARMRCALC24
Function: line_or(bits) line_and(bits)
Use the line_or function to summarize bit vector data over a specified columnar view. Recall that an ARMR data view of a given set of fields shows us only the unique combinations of the specified fields' values. Redundancies are not shown. Nevertheless, we do track the clones of each displayed line.
For each displayed line, line_or returns the union of a given bit-vector data field over the line's clones. The current function is very much like the line_sum, except that we are dealing here with a vector union rather that a numerical sum!
For our example we continue with the example which was started in the previous document (set_bit example). In the previous example we derived single element bit vectors; i.e. the view we are about to use shows the following clones as a function of row:
Row subsets for records with flag=x
-------------------------------------------------
input: row bits clones output: row bits
--- ----------- --- -----
r1 01000 r1 01010
00010 01010
r2 00100 r2 00101
00001 00101
r3 01000 r3 01000
r4 00100 r4 00101
00001 00101
r5 01000 r5 01010
00010 01010
The following calc code produces the outputs shown above:
;
; form the patterns
;
columns= row
calc_start= flagX
line_or(bits)
calc_end=
end=
Finally, we add an additional calc cycle in order to derive the grouping index which collects rows with like description combinations:
;
; define group
;
columns= bits
calc_start= flagX
group=prev(group)+1
calc_end=
end=
The above calc process is as follows:
input: bits row clones output: bits group row clones
----- ----------- ----- ----- -----------
00101 r2 00101 1 r2
r4 r4
01000 r3 01000 2 r3
01010 r1 01010 3 r1
r5 r5
The final calc produces the desired grouping as described in the previous document.
Use the line_and function to summarize bit vector data over a specified columnar view.
The line_and is functionally identical to the line_or, except that we form the intersection rather than the union of the bit vectors.
The current function was included in anticipation of future requirements. We do not show you a practical example. Instead we simply show you the result when a series of bit vectors are subjected to the line_and function:
input bit vectors: output bit vector:
------------------ ------------------
010110101 000100100
000100101
000101100