awk comparison

[shaoji]

Hi,
I have a file as follows:

2|
10|
11|
12|
13|
14|
15|
16d|
2|
3|
31d||
4|
5|
6|
60d|
7|
8|
9|

if I run:

awk 'BEGIN{FS=OFS="|"}{if ($1>a ) {print $1 }; a=$1}' ss

I get

2
10
11
12
13
14
15
16d
2
3

So it means 2>16d>15>,,,>10>2.

and if I run

awk 'BEGIN{FS=OFS="|"}{if ($1>=1 && $1<=5 ) { $1 = "pp" } else if ($1>=6 && $1<=10 ) {$1 = "ss"} print $1}' ss

I got

pp
ss
11
12
13
14
15
pp
pp
pp
pp
pp
pp
ss
60d
ss
ss
ss

It means awk treats 15d as between 1 and 5, 31d between 1 and 5, but does not treat 60d between 6 and 16.

Could someone explain this?

Shaoji

 

[vgersh99]

and how about that:
awk 'BEGIN{FS=OFS="|"}{if ($1+0>=1 && $1+0<=5 ) { $1 = "pp" } else if ($1+0>=6 && $1+0<=10 ) {$1 = "ss"} print $1}'

NOTE: all the comparisons are string/ascii comparisons UNLESS the operands are converted to numbers.

vlad
+----------------------------+
| #include<disclaimer.h> |
+----------------------------+

 

[shaoji]

Thanks Vlad,
I know how to solve this problem. I just don't know how the awk makes comparison. The first example shows, in awk
2>16d>10>2, how could this happen? The second example, why awk treats 16d and 60d differently?
Thanks again.
Shaoji

 

[Ygor]

Awk variables can behave like strings or numbers depending on the context. When awk compares a number to a string the number is converted to a string before performing a lexical comparision.

The lexical sequence of your example data would be...
[tt]
1
15d
16
31d
5
6
60d[/tt]

 

[vgersh99]

The first example shows, in awk
2>16d>10>2, how could this happen?

the comparison is done as strings based on ascii codes and the above is correct: '2' is greater than '16d' (i.e. '2' > '1').
And '16d' is greater than '10' (because '6' > '0'). Why '10' > '2' - I don't know - have to look at your code.

The second example, why awk treats 16d and 60d differently?

What do you mean by 'differently'? I the code I posted - they are treated the same (is you convert them to ints).

vlad
+----------------------------+
| #include<disclaimer.h> |
+----------------------------+

 

[shaoji]

Thanks Vlad. You are right. I think the awk's comparison is the following:

If both are numbers, it compares as numbers (not as in "sort" command), so 10>2. If at least one of them is a string, then it comapre as strings (as in "sort" commmand.)
so 16d>100. It can explain 2>16d>100>5.

The second example is my misunderstanding.

Thanks.

Shaoji

 

[shaoji]

Also thanks to Vgor. You are correct and helpful. In awk, it can happen as follows:

7
8
10
10d
2
3
7

Shaoji