integer division! oh the fun :-)

[chris5126]

Hi guys,

Will keep this short and sweet. Trying to do integer devision to *decimial places but i dont think the version of ksh installed on solaris 10 supports this as they would need to be declared as floats. Is this correct?

If there isnt a way of doing it with ksh is there another way to do it

integer rBytes=`kstat -p -c net -n $Interface -s rbytes64 1 2 |\
                tail -1 | awk '{print $2}'`
  integer oBytes=`kstat -p -c net -n $Interface -s obytes64 1 2 |\
                tail -1 | awk '{print $2}'`
  integer read=$(($rBytes/1000000))
  integer write=$(($oBytes/1000000))
  k=k+1
  echo "rBytes=$rBytes oBytes=$oBytes"
  echo "read=$read Mbps write=$write Mbps"

Gives the output:

rBytes=694127 oBytes=769020
read=0 Mbps write=0 Mbps
rBytes=17061 oBytes=141840
read=0 Mbps write=0 Mbps

Whereas it should give:

rBytes=694127 oBytes=769020
read=0.69 Mbps write=0.76 Mbps
rBytes=17061 oBytes=141840
read=0.17 Mbps write=0.14 Mbps

[Salem]

Personally, I would just do awk '{print $2 / 1000000 }' and stop worrying about just how good the maths is in the shell.

Also, awk has printf() as well, so you've got really good control over the format as well.

[chris5126]

ok not to familar with awk had a little play but getting errors. So if i wanted to manipluate the varialbes and then put them into the new varialbles read and write how would i do it. Also would be good if you could show me how to do it to two decimal places.

Interface=`cat $Interfaces | sed -n $k"p"`
  integer rBytes=`kstat -p -c net -n $Interface -s rbytes64 1 2 |\
                tail -1 | awk '{print $2}'`
  awk '{print $rBytes / 100000}'END
  integer oBytes=`kstat -p -c net -n $Interface -s obytes64 1 2 |\
                tail -1 | awk '{print $2}'`
  #integer read=$(($rBytes/1000000))
  #integer write=$(($oBytes/1000000))
  k=k+1
  echo "rBytes=$rBytes oBytes=$oBytes"

[Salem]

> awk '{print $rBytes / 100000}'END
How is that anything like what I posted?

[chris5126]

haha i was trying to put the value into a varailble and was getting frusturtated and trying different things! thanks for your help!

[eggi]

Hey there,

One way to pass variables to awk (so you don't have to use double quotes) is with the -v flag, like

awk -v rBytes1=$rBytes '{......blah,blah,blah...}'

and then you could use the variable rBytes1 in your script just like any other awk variable.

Best wishes,

Mike

[Fest3er]

bc(1) is your friend. It computes to arbitrary precision.

"-e" tells bash to interpret \x sequences; ksh might not need it.

.
  .
  .
  integer read=`echo -e "scale=2\n$rBytes/1000000" | bc`
  integer write=`echo -e "scale=2\n$wBytes/1000000" | bc`
  .
  .
  .