How to compare current record,with next and previous record in awk without using array?

Shell Programming and Scripting
1

Hi! all

can any one tell me how to compare current record of column with next and previous record in awk without using array

my case is like this

input.txt

0 32
1 26 
2 27
3 34
4 26
5 25
6 24
9 23
0 32
1 28
2 15
3 26
4 24
5 25

output needed

0 32
1 26 
2 27
3 34 BAD
4 26
5 25
6 24
9 23
0 32
1 28
2 15 BAD
3 26
4 24
5 25

just try to implement this logic without using array
and it has to consider first column also that is if next record of column 1 is greater than current record

previous+3 < current < next+3 

or

previous-3 > current < next-3

print BAD

being beginner getting many syntax error those who know please help

2

I do not understand what you try to do.
Post real data, and no example data. It would help to understand what you like to do.

3

I think without array it may be little difficult to handle
you could try

getline and  save column

in some variable and then you can compare

something like this

awk '{p=$2;getline;c=$2;getline;n=$2;print p,c,n}'

I am not sure about this:)

but with array you may get output as you expected from following code, I assume you might be trying to detect peak values in file, test with your original file and post the same as Jotne suggested :b:

                           awk ' FNR==NR{
                                         x=NR
                                         for(i=1;i<=NF;i++)
                                         D[x,i]=$i;next
                                        }

                                    END{
                                           for(j=1;j<=x;j++)
                                           if(j!=1)
                                           if(D[j,1]>D[j-1,1] && D[j-1,2]+3 < D[j,2] && D[j,2] > D[j+1,2]+3 || \
                                           D[j-1,2]-3 > D[j,2] && D[j,2] < D[j+1,2]-3)
           
                                                           print D[j,1],D[j,2],"BAD"
                                           else 
                                                           print D[j,1],D[j,2]
                                        }' your file

output

admin@IEEE:~$ sh test.sh 
1 26
2 27
3 34 BAD
4 26
5 25
6 24
9 23
0 32
1 28
2 15 BAD
3 26
4 24
5 25
4

Not clear. Why is neither your record 9 nor second record 0 "BAD"? Please explain your logics in plain English.

5

That is because column 1 has 0
Either a zero or any value which is less than previous would be considered as new data
for example

0 xx  data 1
1 xx  data 1
2 xx  data 1
3 xx  data 1
2 xx data 2
3 xx  data 2
4 xx data 2
5 xx data 2
6 xx data 2

Akshay's code is giving some result but very slow, I am having about 4 GB data file

6

Still not clear why "1 26" (line #2) is not BAD.

  • Not "new data", by your definition.
  • Previous number was 32, so difference is 4.

Can you compile a simple C program? This would be easy to write in C, if you can spell out exactly what you want to do. (And if that does not break the forum rules :D)

It would be interesting to compare the performance of some C program vs some awk solution.

7

Yes hanson44...I am also interested...
Thats why I left $0 records by

j!=1
8

Please revise and explain/rephrase that logics.