*** pgsql/src/backend/commands/analyze.c	2008/11/10 00:49:37	1.128
--- pgsql/src/backend/commands/analyze.c	2008/12/13 19:13:44	1.129
***************
*** 8,14 ****
   *
   *
   * IDENTIFICATION
!  *	  $PostgreSQL: pgsql/src/backend/commands/analyze.c,v 1.127 2008/11/02 01:45:27 tgl Exp $
   *
   *-------------------------------------------------------------------------
   */
--- 8,14 ----
   *
   *
   * IDENTIFICATION
!  *	  $PostgreSQL: pgsql/src/backend/commands/analyze.c,v 1.128 2008/11/10 00:49:37 tgl Exp $
   *
   *-------------------------------------------------------------------------
   */
*************** typedef struct AnlIndexData
*** 67,73 ****
  
  
  /* Default statistics target (GUC parameter) */
! int			default_statistics_target = 10;
  
  /* A few variables that don't seem worth passing around as parameters */
  static int	elevel = -1;
--- 67,73 ----
  
  
  /* Default statistics target (GUC parameter) */
! int			default_statistics_target = 100;
  
  /* A few variables that don't seem worth passing around as parameters */
  static int	elevel = -1;
*************** std_typanalyze(VacAttrStats *stats)
*** 1531,1540 ****
  		 * error in bin size f, and error probability gamma, the minimum
  		 * random sample size is
  		 *		r = 4 * k * ln(2*n/gamma) / f^2
! 		 * Taking f = 0.5, gamma = 0.01, n = 1 million rows, we obtain
  		 *		r = 305.82 * k
  		 * Note that because of the log function, the dependence on n is
! 		 * quite weak; even at n = 1 billion, a 300*k sample gives <= 0.59
  		 * bin size error with probability 0.99.  So there's no real need to
  		 * scale for n, which is a good thing because we don't necessarily
  		 * know it at this point.
--- 1531,1540 ----
  		 * error in bin size f, and error probability gamma, the minimum
  		 * random sample size is
  		 *		r = 4 * k * ln(2*n/gamma) / f^2
! 		 * Taking f = 0.5, gamma = 0.01, n = 10^6 rows, we obtain
  		 *		r = 305.82 * k
  		 * Note that because of the log function, the dependence on n is
! 		 * quite weak; even at n = 10^12, a 300*k sample gives <= 0.66
  		 * bin size error with probability 0.99.  So there's no real need to
  		 * scale for n, which is a good thing because we don't necessarily
  		 * know it at this point.