awk Search Array Element Return Index

Shell Programming and Scripting
1

Can you search AWK array elements and return each index value for that element.

For example an array named car would have index make and element engine. I want to return all makes with engine size 1.6.

Array woulld look like this:

BMW 1.6
BMW 2.0
BMW 2.5
AUDI 1.8
AUDI 1.6

Output I want is:

BMW
AUDI
2

Try this:

awk '{if ($2 == 1.6 ) { print $1} }' inputFile

Or this

awk '$2 ~ /1\.6/ { print $1}' inputFile
3

I don't think you can have that array as the index needs to be unique. One possible index could be "make & engine". Then you could try sth like (untested)

for (i in array) {split (i, M); X[M[1]]} for (i in X) print i}
4

Agreed - stupid example. Was trying to make it simple.

Real world - we have two arrays (source & comparison). What I want is for each source index element value return all the element matching comparison index values.

Source:
102030 BEE77DA5BE9F24FBE044002128B2E77C

Comparison:
102030 BEE77DA5BE9F24FBE044002128B2E77C
102040 BEE77DA5BE9F24FBE044002128B2E77C
102050 AEE77123402128B2EF24FBEE77D28B2E
102080 BEE77DA5BE9F24FBE044002128B2E77C

Results:
Source: 102030
Result: 102040
Result: 102080

5

Here is an example performing array value comparison.

awk '
        BEGIN {
                SRC["102030"] = "BEE77DA5BE9F24FBE044002128B2E77C"
                CMP["102030"] = "BEE77DA5BE9F24FBE044002128B2E77C"
                CMP["102040"] = "BEE77DA5BE9F24FBE044002128B2E77C"
                CMP["102050"] = "AEE77123402128B2EF24FBEE77D28B2E"
                CMP["102080"] = "BEE77DA5BE9F24FBE044002128B2E77C"
        }
        END {
                print "Results:"
                for ( v in SRC )
                {
                        for ( k in CMP )
                        {
                                if ( k == v )
                                        print "Source: " k
                                else if ( CMP[k] == SRC[v] )
                                        print "Result: " k
                        }
                }
        }
' /dev/null

Modify as per you requirement.

6

Thank you. Was struggling with syntax.

If you don't initialise v it means loops through all records.

for ( v in SRC )
7

Yes, there is no need to initialize variable: v

It is a special kind of for statement for scanning an array.

8

Struggling to obtain the correct output format.

What I want is (problem name and hash values should line up):

Source_Name Problem_Name Hash
102030         102080           BEE77DA5BE9F24FBE044002128B2E77C
                   102040           BEE77DA5BE9F24FBE044002128B2E77C

Amended script:

nawk -F= '
        BEGIN {
                SRC["102030"] = "BEE77DA5BE9F24FBE044002128B2E77C"
                CMP["102030"] = "BEE77DA5BE9F24FBE044002128B2E77C"
                CMP["102040"] = "BEE77DA5BE9F24FBE044002128B2E77C"
                CMP["102050"] = "AEE77123402128B2EF24FBEE77D28B2E"
                CMP["102080"] = "BEE77DA5BE9F24FBE044002128B2E77C"
        }
        END {
                printf "Source_Name" " Problem_Name Hash\n"
                for ( v in SRC )
                {
                        for ( k in CMP )
                        {
                               if ( k == v )
                                        printf k
                               else if ( CMP[k] == SRC[v] )
                                        printf "%+13s\n",k CMP[k]
                        }
                }
        }
' /dev/null

Output:

Source_Name Problem_Name Hash
102030102080BEE77DA5BE9F24FBE044002128B2E77C
102040BEE77DA5BE9F24FBE044002128B2E77C
9

modify

else if ( CMP[k] == SRC[v] )
                                        printf "%+13s\t%s\n",k,CMP[k]

Separators were missing

10

Almost, 102040 should line up with 102080.

Source_Name Problem_Name Hash
102030      102080      BEE77DA5BE9F24FBE044002128B2E77C
      102040    BEE77DA5BE9F24FBE044002128B2E77C

Maybe I'm going about it the wrong way. How would I set a fixed starting point for:

printf "%+13s\t%s\n",k,CMP[k]
11

Try

        for ( k in CMP )
                {printf "%13s\t", (k==v)?k:""
                 if ( CMP[k] == SRC[v] )
                        printf "%+13s\t%s\n",k,CMP[k]
                }
12

Not quite but I can see where you're going. I'll play about with it a bit.

Source_Name Problem_Name Hash
      102030          102030    BEE77DA5BE9F24FBE044002128B2E77C
                      102080    BEE77DA5BE9F24FBE044002128B2E77C
                                      102040    BEE77DA5BE9F24FBE044002128B2E77C

---------- Post updated at 02:38 PM ---------- Previous update was at 01:59 PM ----------

Sod it. Made it easier.

for ( k in CMP )
                        {
                              if ( CMP[k] == SRC[v] )
                                  printf ("%-12s%-13s%s\n",v,k,CMP[k]) 
                        }