Awk - Simple but powerful

Everything happened today while I was trying to visualize a set of items and its frequencies. Usually, till date, I had converted it to CSVs and then loaded into excel(in my office system!!). Everyone could easily interpret the values in the dataset this way. But when I needed to analyse this data alone, it became a little inconvenient for me. Thus I thought of writing a script which would help me to visualize this set of data in ASCII format. I know that there are many tools out there which would really help me do it through command prompt. But here I was in a situation where I would not be able to install anything new. Thus AWK came to the rescue…

Before going further into what I had done, I wanted to take some time to see what I had understood regarding the language from one of its creators.

As Brian Kernighan, said in one of his lectures, which I accidentally saw last week, writing a language for a specific task can help in solving that problem very well than writing a general purpose language every time which will include all the unwanted and complex constructs. These undesirable features in a general purpose language will only slow down the process of problem solving rather than helping to solve it efficiently.

AWK has solved the problem of text processing very efficiently, with many of the commonly used program patterns coded into the language itself. Thus we don’t have to go behind solving all the problems that is in front of us and just focus on the core things that are needed.

The need to parse the files line by line and execute some kind of common code pattern is very usual in text processing tasks. This is very neatly implemented in the language. The programmer just has to write only what the repeating code pattern is.

Each line can be broken down into columns - this by default uses space as a field separator. The field separators can also be changed.

As Mr.Kernighan said, if there is only one class that you can teach him data structures, then teach him associated arrays. This data structure has been used in AWK. It helps in maintaining a status of everything that we had done during our processing stage, and this status list can be utilized during the final consolidating stage.

Now coming back to the point what I was talking about…
Below given is the script that I had written to take the input as a list of two member set in a file, where the first member is the count and the second element is the identifier for each item. The script will print out the identifier and the horizontal bar graph for the entire list of items. I had to do a normalization for the values in count field, since the range of values may not be that predictable.

awk 'BEGIN \
     { min = -1 } \
     { if (min == -1) { min = $1} else if (min > $1) {min = $1}; \
       if (max < $1) max = $1; normal_count[$2] = $1; \
       sum += $1 \
     } \
     END \
     { diff = max - min; \
       for (i in normal_count) { normal_count[i] = (normal_count[i] - min)*100/diff; printf "%3d ", i; \
         for ( j=0; j<normal_count[i]; j++ ) \
         { printf "="; } print ""; \
         } \
     }' input | less

NOTE1: Thank you Ashwin for reading the first draft of this post.

NOTE2: Mostly all the points mentioned are taken directly from the lecture mentioned above.

NOTE3: I may have to rewrite the article.