Proc report file size in GBs

[sathya66]

Hi All,

we have coulm called size and that is defined by byites but we need to convert that varaible into GBs/TBs. and total size of all files.

can you advice on this.

proc report data=test nofs; 
 title1 "Simple Report";
 column FILE  SIZE ; 
 define FILE / display;
 define SIZE / analysis sum; 
 rbreak after /  summarize;
 compute SIZE;
size=size/1024;
 endcomp;
 run; 

[PaigeMiller]

Please tell us what is not working properly with the code you show.

[sathya66]

Hi

sorry I was not crear.

size column is in bytes so we want to convert that into GBs/Tbs

ex:sample data in bytes.

need an output like below as bytes will be caliclated like size=(((size)/1024)/1024) to get in  GBs

file	size(in GBs)
aa	0.000977
bb	0.00193
cc	0.001812
ff	1.3125
hh	20.8125
total

22.12972

data test;
infile cards;
input file $ size;
cards;
aa 1024
bb 2024
cc 1900
ff 1376256
hh 21823488
;
run;



proc report data=test nofs; 
 title1 "Simple Report";
 column FILE  SIZE ; 
 define FILE / display;
 define SIZE / analysis sum; 
 rbreak after /  summarize;
 compute SIZE;
size=size/1024;
 endcomp;
 run; 

[PaigeMiller]

---

@sathya66 wrote:

Hi

sorry I was not crear.

size column is in bytes so we want to convert that into GBs/Tbs

ex:sample data in bytes.

need an output like below as bytes will be caliclated like size=(((size)/1024)/1024) to get in  GBs

file	size(in GBs)
aa	0.000977
bb	0.00193
cc	0.001812
ff	1.3125
hh	20.8125
total

22.12972

---

It's still not clear to me. You haven't told me what is wrong with the program or the output you get from the program, other than the very obvious you want a different formula for SIZE in the compute block, and you state above the correct formula.

[Kurt_Bremser]

I played around with the format to make it look a little better:

proc format;
picture sizes
  0-10240 = "00,009"
  10240-10485760 = "00,009.00K" (mult=0.09765625)
  10485761 - 10737418240 = "00,009.00M" (mult=9.5367431640625E-5)
  10737418241 - 10995116277760 = "00,009.00G" (mult=9.3132257461548E-8)
  10995116277761 - high = "0,000,009.00T" (mult=9.0949470177292E-11)
;
run;

[Kurt_Bremser · Accepted solution]

Use an adaptive format:

data test;
input file $ size;
datalines;
A 2000
B 20000
C 200000
D 2000000
E 20000000
F 200000000
G 2000000000
H 20000000000
I 200000000000
J 2000000000000
K 20000000000000
;

proc format;
picture sizes
  0-10240 = "00009"
  10240-10485760 = "00009K" (mult=0.0009765625)
  10485761 - 10737418240 = "00009M" (mult=9.5367431640625E-7)
  10737418241 - 10995116277760 = "00009G" (mult=9.3132257461548E-10)
  10995116277761 - high = "0000009T" (mult=9.0949470177292E-13)
;
run;

proc report data=test nofs; 
 title1 "Simple Report";
 column FILE  SIZE ; 
 define FILE / display;
 define SIZE / analysis sum format=sizes.; 
 rbreak after /  summarize;
run;

You can tweak the cutoff values, and the picture to include digits after the decimal point.