Job_Name Predecessor
A NULL
B1 A
B2 A
B3 B1
C B3
C B2
So based on these i have to generate the level Number
What i mean is
Let A be level 1
for B1 to happen it should have done A
so B1 level is A+1 = 1+1 = 2
and same for B2 it should have done A
so B2 level is A+1 = 1+1 = 2
for B3 to happen it should have done B1
so B3 level is B2+1 = 2+1 = 3
but for C to complete it should have done B2 and B3
so C is max(B2,B3)+1
C=max(2,3)+1=4
$ cat data1
A NULL
B1 A
B2 A
B3 B1
C B3
C B2
$ awk '{r[NR]=$0 ; k[NR]=$1 ; if(v[$1]<1+v[$2]) v[$1]=1+v[$2] } END { for (n=1; n<=NR; n++) print r[n] , v[k[n]]}' data1
A NULL 1
B1 A 2
B2 A 2
B3 B1 3
C B3 4
C B2 4
$
This will never work unless you sort the input so that each job comes after all of its predecessors. The data pairs you have represent a directed graph. Provided that that the directed graph is acyclic, such a sort should be possible and this is called a topological sort. Lucky for you, Unix has a program called tsort to do this. Try:
But bear in mind that with a complex directed graph there will probably be several different ways to assign your level numbers. This will find one of them.
Put the script back the way I wrote it. You cannot introduce random changes to my script and expect it to continue to work. I will test my script, not yours. My script with E in the data:
$ cat datae
A 0
B 0
C 0
D 0
E A
E B
E C
Z H
Z D
F E
G E
H E
I G
$ ./level datae
A 0 1
B 0 1
C 0 1
D 0 1
E A 2
E B 2
E C 2
F E 3
G E 3
H E 3
I G 4
Z D 4
Z H 4
$
And my script with P in the data:
$ cat datap
A 0
B 0
C 0
D 0
P A
P B
P C
Z H
Z D
F P
G P
H P
I G
$ ./level datap
A 0 1
B 0 1
C 0 1
P A 2
P B 2
P C 2
D 0 1
F P 3
G P 3
H P 3
I G 4
Z D 4
Z H 4
$
The only difference is that the "D 0 1" moved below the E/P lines. That is an artifact of tsort, there are several possible ways to order the data and tsort displays one of them. If this bothers you, use a final sort to sort the output.
but i dont know y it shows as all 1's in the 3rd column...
and then
that gets me the Job name and the interger ... doestnt get me the job and predecessor as expected...
thats y i thought that i might be
instead of
..
the join of temp1 and - doesnt get me the job and predeessor...
Please throw some light on it...
thanks for all the help...
if i am right
sort $1 > temp1
tsort temp1 --- does the topological sort
sed '$d' --- deletes the last line which is 0 in our case
nl --- appends line numbers to each line
sort -nr --- sort the file in reverse based on the line numbers given by previous nl command
cut -f2 --- gets the job name back ...
nl --- again a line number appending
awk '{print $2, $1}'--- gets the job name followed by line number to be printed
sort --- sort on the job name
join temp1 - --- ??????? dont know exactly what it does.... { i thought it joins the Job name to its predecessor}
sort -k 3n ---- sort on the 3rd field
awk '{print $1,$2}' -- prints the first two columns... (in this case i thought the first 2 colums would be the job name, line number )
the next login i flaw less i think the tsort is fine....
awk '{r[NR]=$0 ; k[NR]=$1 ; if(v[$1]<1+v[$2]) v[$1]=1+v[$2] } END { for (n=1; n<=NR; n++) print r[n] , v[k[n]]}'
I am counting on You people only....
Please dont desert me.....
join does a "join" operation on two files, where "join" is the relational database join. Since I did not specify otherwise, the first field of both files is the key field and the files must be sorted on the field. Example:
$ cat one
dog puppy
cat kitty
$ cat two
dog fido
dog rover
dog spot
cat fluffy
cat tiger
cat felix
$ join one two
dog puppy fido
dog puppy rover
dog puppy spot
cat kitty fluffy
cat kitty tiger
cat kitty felix
The first output field is the key field for both files. Then comes the data fields for the first file and then come the data fields for the second file. This is how join is supposed to work. You seem to be saying that your join is giving the key field, then data field for the 2nd file, and then the data field for the first file. It's not supposed to do that. But I can't test on AIX. If it is backwards, then your awk change should compensate. But if you still get wrong output, something else must not be right. Not sure what...
