It then runs through the array outputting the required column from all the stored lines. The first time the search word is found, it stores the field number it was in. This will be fine as long as the search string is in maybe the first 10,000 lines. This means it stores the minimum data necessary to make it able to read the file just once. This reads the input file once, and it stores all the initial lines in an array, up to and including the first line that has a field exactly equal to the search word. There now, you not only know exactly what you are trying to do, but it almost looks like we have a procedure emerging. If the search word column is now well-defined, list only the corresponding columns in all lines in the file, including those before the search word was found.ĭo not list the column from lines that do not have at least that many columns present. If the search word is not found, no output is produced. The column number used in the first such line containing this value is used for the whole file. Look through the file until there is a field containing exactly 'xxxx' where xxxx is a parameter to the function. How about: The file is defined in columns what are white-space separated, any number of consecutive tabs and spaces constituting one separator. You really need to sharpen up the specification, and think through how can it can be made fool-proof, before you seek a solution. Do you switch to outputting column 7 from there on? Do you want to go back to the start of the file and output column 7 from the top this time. Surely you can see that needs two passes through the file, or to hold the whole file in an array so you can process it twice.Īlso, what happens when you get to line 345 and 'Size' turns up in column 7.

The hard bit is that, having found 'Size' on the second line, you expect to be able to magically reprocess the first line to get the word 'Signal'.

Awk find word in file