26160 - Execute a 'chained' lookup using a hash table

Sample 26160: Execute a 'chained' lookup using a hash table

Load a data set view into a hash table. Use the hash table to track the links that have contributed to a 'chain' and to reset the variable that is used to trigger the next look-up in the 'chain'.

Read observations from a data set. Use the FIND() method to continuously search for 'links' in the hash table, until the look-up fails. This indicates you are at the end of the chain. Create a new variable of each 'link' in the 'chain' using CATX.

/* Create starting data */
data transactions;
  input id $ new_id $;
datalines;
ab1234 cf3245
klh3409 we93476
cf3245 lk23489
34pwk32 whlw09
we93476 34pwk32
fds0987 welz345
;

/* When using the DATASET: argument tag, all variables in the hash table must  */
/* exist in the data set.  In this case, we need a flag variable that does not */
/* exist on the current version of TRANSACTIONS.  A quick way around this is   */
/* to create a VIEW of TRANSACTIONS that does contain the flag variable        */
/* ALREADY_USED.                                                               */

/* Create a view of the desired data set plus the additional variable.         */
data transactions_vw/view=transactions_vw;
  retain already_used ' ';
  set transactions;
run;

data links;
  keep chain start end;
  length chain $100 start $8 end $8;

  /* On the first iteration of the DATA step, create the hash table H. */
  if _n_=1 then do;
    /* Create hash table H to hold unique ID values from TRANSACTIONS.  */
    /* The DATASET: argument tag loads the view TRANSACTIONS_VW into H. */
    declare hash h(dataset: 'transactions_vw');
    h.definekey('id');
    h.definedata('id','new_id','already_used');
    h.definedone();
  end;

  /* Read values for ID from TRANSACTIONS.  Use the FIND method to locate a   */
  /* match for ID's value in the hash table H.  Grab the NEW_ID value from    */
  /* H and assign its value into ID.  Repeat the look up with the FIND method */
  /* to determine if the NEW_ID value assigned into ID has a match.  If so,   */
  /* you are in a chain.  Continue 'chaining' until no match is found.        */
  do until(done);
    set transactions end=done;
    test=h.find();
    if test=0 and already_used='y' then continue;
    /* Assign NEW_ID from TRANSACTIONS into ID, for the look up.  Create a      */
    /* new variable CHAIN to track each 'link' in the chain.  This is optional. */
    /* To trap only the 'end points' of a chain, create 2 new variables, START  */
    /* and END.                                                                 */
    chain=catx(' ',id, new_id);
    start=id;
    end=new_id;
    /* NEW_ID will become the new value to match in the hash table. */
    id=new_id;

    do until(rc ne 0);  /* Keep 'chaining' until no match is found.  Be sure to */
                        /* know your data.  If you have unexpected overlap in   */
                        /* the ranges, such as one 'link' pointing to a second  */
                        /* link that points back to the first, you could enter  */
                        /* an infinite loop.                                    */
      rc=h.find();
      if rc=0 then do;
        /* Mark the item in the hash table as 'used'.  This flag will keep */
        /* individual 'links' from being output as separate 'chains' of    */
        /* their own.                                                      */
        already_used='y';
        h.replace();
        /* Tack NEW_ID onto the end of CHAIN.  Reassign END and ID accordingly. */
        chain=catx(' ',chain,new_id);
        end=new_id;
        id=new_id;
      end;
      /* If RC is not zero, no match was found, so output the observation. */
      else output;
    end; /* inner DO UNTIL */
  end; /* outer DO UNTIL */
run;

proc print;
run;

These sample files and code examples are provided by SAS Institute Inc. "as is" without warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability and fitness for a particular purpose. Recipients acknowledge and agree that SAS Institute shall not be liable for any damages whatsoever arising out of their use of this material. In addition, SAS Institute will provide no support for the materials contained herein.

Obs    chain                              start       end

 1     ab1234 cf3245 lk23489             ab1234     lk23489
 2     klh3409 we93476 34pwk32 whlw09    klh3409    whlw09
 3     fds0987 welz345                   fds0987    welz345

Type:	Sample
Topic:	SAS Reference ==> DATA Step SAS Reference ==> Component Objects SAS Reference ==> Component Objects ==> hash object SAS Reference ==> Component Objects ==> hash object ==> FIND SAS Reference ==> Functions ==> Character ==> CATX

Date Modified:	2007-05-04 03:03:05
Date Created:	2007-05-04 03:03:05

Product Family	Product	Host	SAS Release
Product Family	Product	Host	Starting	Ending
SAS System	Base SAS	z/OS	9.1 TS1M0
		Microsoft® Windows® for 64-Bit Itanium-based Systems	9.1 TS1M0
		Microsoft Windows Server 2003 Datacenter 64-bit Edition	9.1 TS1M0
		Microsoft Windows Server 2003 Enterprise 64-bit Edition	9.1 TS1M0
		Microsoft Windows 2000 Advanced Server	9.1 TS1M0
		Microsoft Windows 2000 Datacenter Server	9.1 TS1M0
		Microsoft Windows 2000 Server	9.1 TS1M0
		Microsoft Windows 2000 Professional	9.1 TS1M0
		Microsoft Windows NT Workstation	9.1 TS1M0
		Microsoft Windows Server 2003 Datacenter Edition	9.1 TS1M0
		Microsoft Windows Server 2003 Enterprise Edition	9.1 TS1M0
		Microsoft Windows Server 2003 Standard Edition	9.1 TS1M0
		Microsoft Windows XP Professional	9.1 TS1M0
		64-bit Enabled AIX	9.1 TS1M0
		64-bit Enabled HP-UX	9.1 TS1M0
		64-bit Enabled Solaris	9.1 TS1M0
		HP-UX IPF	9.1 TS1M0
		Linux	9.1 TS1M0
		OpenVMS Alpha	9.1 TS1M0
		Tru64 UNIX	9.1 TS1M0

Support

Sample 26160: Execute a 'chained' lookup using a hash table

Operating System and Release Information