# awk, max value, array, row

**URL:** <https://community.unix.com/t/awk-max-value-array-row/313628>\
**Category:** Shell Programming and Scripting\
**Created:** [July 16, 2012, 2:42pm UTC](https://community.unix.com/t/awk-max-value-array-row/313628 "2012-07-16T14:42:35Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![yifangt](https://community.unix.com/letter_avatar/yifangt/32/6_22734026db12803ebb3e059622ac3534.png) [@yifangt](https://community.unix.com/u/yifangt)\
**Post date:** [July 16, 2012, 2:42pm UTC](https://community.unix.com/t/awk-max-value-array-row/313628/1 "2012-07-16T14:42:35Z")

</div>

Hello:  
I want to print out the entire row with max value in column 3 based on column 2. Input file is millions rows. test.dat:

```nohighlight
Contig1 lcl|1DL 111 155 265 27
Contig2 lcl|1DS 100 73 172 100
Contig3 lcl|1DL 140 698 837 140
Contig3 lcl|6DS 107 1488 1594 1
Contig5 lcl|6DL 193 59 251 374
Contig5 lcl|4DS 119 1 119 119
Contig5 lcl|6DL 107 145 251 596
Contig6 lcl|6DS 153 90 242 674
Contig7 lcl|4DL 103 913 1015 6590
Contig7 lcl|6DL 107 1016 1122 1152
Contig8 lcl|6DS 291 2700 2990 291
Contig8 lcl|4DS 279 2594 2872 279
Contig8 lcl|6DS 244 3711 3954 1
Contig8 lcl|6DS 159 3796 3954 1
Contig8 lcl|6DL 194 3237 3430 194
Contig8 lcl|1DS 109 4069 4177 269

```

I first tried:

```nohighlight
awk '{if(! ($2 in a)) a[$2]=$3; else if($3 > a[$2]) a[$2]=$3; max[$2]=$0} END {for (i in max) print i, a}' test.dat 

```

and the output is:

```nohighlight
lcl|4DL 103
lcl|4DS 279
lcl|6DL 194
lcl|6DS 291
lcl|1DL 140
lcl|1DS 109

```

As I want to print out the whole row of the max values of each item, then I tried:

```nohighlight
awk '{if(! ($2 in a)) a[$2]=$3; else if($3 > a[$2]) a[$2]=$3; max[$2]=$0} END {for (i in max) print max}' test.dat

```

and the output is:

```nohighlight
Contig7 lcl|4DL 103 913 1015 6590
Contig8 lcl|4DS 279 2594 2872 279
Contig8 lcl|6DL 194 3237 3430 194
Contig8 lcl|6DS 159 3796 3954 1
Contig3 lcl|1DL 140 698 837 140
Contig8 lcl|1DS 109 4069 4177 269

```

Obviously I had something wrong with the second script. I am very nervous with the second script for millions of rows, but could not figure it out myself. Thanks in advance!  
YT

---

<div class="post-metadata">

**Author:** ![alister](https://community.unix.com/letter_avatar/alister/32/6_22734026db12803ebb3e059622ac3534.png) [@alister](https://community.unix.com/u/alister)\
**Post date:** [July 16, 2012, 2:59pm UTC](https://community.unix.com/t/awk-max-value-array-row/313628/2 "2012-07-16T14:59:35Z")

</div>

The max assignment is outside the if-else statement, so it always occurs.

Regards,  
Alister

---------- Post updated at 02:59 PM ---------- Previous update was at 02:55 PM ----------

Also, you don't really need to bother with checking for the existence of an array member in awk. If it does not exist, it's treated as a zero or an empty string (depends on the context). Your script could be simplified a bit:

```nohighlight
awk '$3 > a[$2] {a[$2]=$3; max[$2]=$0} END {for (i in max) print max}' test.dat

```

Regards,  
Alister