topic Re: some form of masking in SAS Programming

some form of masking

HeatherNewton — Wed, 10 May 2023 07:30:44 GMT

data cust;
input cust_id $8.;
datalines;
E4349567
C4127658
C6267260
E5211402 
K4729617
G6646262
;
run;

PROC PRINT DATA=CUST;
RUN;

data test2;
 set cust;
 string=COLLATE(1,256);
 high4=substr(string,55);
 cust_id=translate(cust_id, high4, string);
run;

proc print data=test2;
run;

Can anyone let me know why I only have 1 entry in my output datset test2, I should have same number as my input datatset cust, right?

Re: some form of masking

PeterClemmensen — Wed, 10 May 2023 07:44:35 GMT

When I run your code, I have 6 observations in both.

Re: some form of masking

Kurt_Bremser — Wed, 10 May 2023 07:49:18 GMT

When something unexpected happens, always (as in ALWAYS) post your complete log.

Re: some form of masking

HeatherNewton — Fri, 08 Sep 2023 10:35:18 GMT

data test2;
 set cust;
 string=COLLATE(1,256);
 high4=substr(string,55);
 cust_id=translate(cust_id, high4, string);
run;

does these create binary code? they look like symbols e.g.

^@000^0^0

Re: some form of masking

Kurt_Bremser — Fri, 08 Sep 2023 11:26:43 GMT

By shifting characters "up" by 55, you will eventually get results in the range of 128-182. Within this range, you have bytes which initiate a multi-byte UTF sequence, or are only allowed as UTF continuation bytes.

If you need a reversible process which creates readable results and makes it at least a little hard to reverse for a third person, create a lookup table with values created by a random function. How you create such depends on the structure of the values to be encoded.

Re: some form of masking

HeatherNewton — Sat, 09 Sep 2023 13:22:23 GMT

can I use encoding=utf8 to display it?

I used encoding=utf8 to convert the file to csv and use codepage=1208 (which is also for utf8)

to load into db2

however the content splited up into two rows with a cache return between

so e.g. ^@000^@^@

before ^@000 ^@^@

does it work if I use double quote to enclose it before converting into csv

but I tired it keep giving me """"""""""""

Re: some form of masking

Kurt_Bremser — Sat, 09 Sep 2023 17:20:04 GMT

If you want something readable which will reliably load into another software, you need a completely different approach. This crude "shift by 55" is clearly not the way to go.

Re: some form of masking

Patrick — Sun, 10 Sep 2023 01:28:30 GMT

@HeatherNewton The very similar forum discussion here provides multiple options how you could approach this.

Re: some form of masking

HeatherNewton — Mon, 13 Nov 2023 08:02:20 GMT

data test2;
 set cust;
 string=COLLATE(1,256);
 high4=substr(string,55);
 cust_id=translate(cust_id, high4, string);
run;

what would be the code in sas to reverse the above?

Re: some form of masking

Patrick — Mon, 13 Nov 2023 08:54:02 GMT

Just reverse the from & to in the translate function.

data test2;
  set cust;
  string=COLLATE(1,256);
  high4=substr(string,55);
  cust_id_out_1=translate(cust_id, high4, string);
  cust_id_out_2=translate(cust_id_out_1,string,high4);
run;

Re: some form of masking

HeatherNewton — Mon, 13 Nov 2023 10:08:29 GMT

data custinfo;
input cust_id $8.;
datalines;
E4349567
C4127658
C6267260
E5211402 
K4729617
G6646262
;
run;

data test1;
 set custinfo;
 string=COLLATE(1,256);
 high4=substr(string,55);
 cust_id_out_1=translate(cust_id, high4, string);
 cust_id_out_2=translate(cust_id_out_1,string,high4);
run;

proc print data=custinfo;
run;
proc print data=test1;
run;

I tried but cannot see cust_id_out_1 or cust_id_out_2, but no error though

Re: some form of masking

Quentin — Mon, 13 Nov 2023 12:02:04 GMT

Try starting a fresh SAS session and run it again. The code works for me.

Re: some form of masking

Patrick — Mon, 13 Nov 2023 23:38:59 GMT

@HeatherNewton Works for me without any issues once I apply format $hex<n>. to print variables with values that can contain control characters and the like.

data custinfo;
  input cust_id $8.;
  datalines;
E4349567
C4127658
C6267260
E5211402 
K4729617
G6646262
;
run;

data test1;
  set custinfo;
  string=COLLATE(1,256);
  high4=substr(string,55);
  cust_id_out_1=translate(cust_id, high4, string);
  cust_id_out_2=translate(cust_id_out_1,string,high4);
  format string $hex512. high4 $hex110. cust_id_out_1 $hex16.;
run;

proc print data=custinfo;
run;

proc print data=test1;
run;

Btw: If you really want to create and store a list of the first 256 characters in the collating sequence then you need to explicitly define the length for variable string as else it will default to a length of only $200 in your code.

You also need to start counting at 0 and not 1.

And... if you want to ensure consistent results for running in any SAS environment then you need also to care about the Collating Sequence and define it explicitly.

I'd also consider to create a list of characters for variable string that only contains printable characters and that certainly don't contain control characters.