/*<pre><b>
/ Program   : clength.sas
/ Version   : 1.0
/ Author    : Roland Rashleigh-Berry
/ Date      : 29-Mar-2007
/ Purpose   : To create a length statement to unify character lengths in a list
/             of data sets to the maximum variable length.
/ SubMacros : %words %nobs
/ Notes     : This is not a function-style macro. A length statement will be 
/             generated in the form "length cvar1 $ 5 cvar2 $ 12" BUT ONLY IF
/             THERE IS AN INCONSISTENCY IN THE INPUT DATASETS. Otherwise it will
/             be blank. Names, labels and other attributes will be taken from
/             the first data set in the list. The length statement string is
/             written out to a global macro variable which can then be resolved
/             in a later data step.
/ Usage     : %clength(ds1 ds2 ds3);
/             data all;
/               &_clength_;
/               set ds1 ds2 ds3;
/             run;
/===============================================================================
/ PARAMETERS:
/-------name------- -------------------------description------------------------
/ dsets             (pos) Input datasets
/ globvar=_clength_ Name of global macro variable to write length string to
/===============================================================================
/ AMENDMENT HISTORY:
/ init --date-- mod-id ----------------------description------------------------
/ rrb  29Mar07         Put out "macro called" message plus header tidy
/===============================================================================
/ This is public domain software. No guarantee as to suitability or accuracy is
/ given or implied. User uses this code entirely at their own risk.
/=============================================================================*/

%put MACRO CALLED: clength v1.0;

%macro clength(dsets,globvar=_clength_);

%local i w;
%let w=%words(&dsets);

%global &globvar;
%let &globvar=;

%do i=1 %to &w;
proc contents data=%scan(&dsets,&i,%str( )) noprint
out=_clen&i(keep=name type length where=(type=2));
data _clen&i;
  retain seq &i;
  length ucname $ 32;
  set _clen&i;
  ucname=upcase(name);
  drop type;
run;
%end;

*- bring all the data sets together -;
data _clenall;
  set
  %do i=1 %to &w;
    _clen&i
  %end;
  ;
run;

*- sort ready to get first form of variable name -;
proc sort data=_clenall;
  by ucname seq;
run;

*- first form of variable name encountered -;
data _clenf;
  set _clenall(keep=ucname name);
  by ucname;
  if first.ucname;
  rename name=fname;
run;

*- merge first form of name in with rest -;
data _clenall(keep=fname length);
  merge _clenf _clenall;
  by ucname;
run;

*- get rid of duplicate lengths -;
proc sort nodupkey data=_clenall;
  by fname length;
run;

*- sort in descending length order -;
proc sort data=_clenall;
  by fname descending length;
run;

*- we only want the one with the longest length where there is a clash -;
data _clenall;
  set _clenall;
  by fname;
  if first.fname and not last.fname then output;
run;

%if %nobs(_clenall) %then %do;
*- gemerate the length statement and output to global macro variable -;
data _null_;
  length str $ 32767;
  retain str 'length';
  set _clenall end=last;
  str=trim(str)||' '||trim(fname)||' $ '||compress(put(length,5.));
  if last then call symput("&globvar",trim(str));
run;
%end;

*- tidy up -;
proc datasets nolist;
  delete _clen:;
run;
quit;

%mend;