c语言快速排序（C language quick sort）.docVIP

c语言快速排序（C language quick sort） 9.3.2 quick sort In the discussion of bubble sort, because the scanning process only to the two adjacent elements are compared, so only in the exchange of two adjacent elements to eliminate a reverse. If by two (non adjacent elements) to exchange, can eliminate multiple decommitment records, will greatly accelerate the speed of sorting. Fast sorting method is to think through an exchange and eliminate multiple reverse. The basic idea is: quick sort, select a record from the records to be sorted in the sequence (usually select the first record), whose key is K1, then the rest will be less than K1 keyword records moved to the front, and the key is greater than K1 records will be moved to the back, the sequence is divided into two sub sequence records table, finally keywords for K1 records are inserted into the position of the boundaries. We will call this process a quick sort. Through a partition, with keywords for K1 records as the dividing line, will be divided into two sub sequence table, and all records before the face keyword in the table are not larger than K1, then all the records in the table face keywords are not less than K1. The segmented sub table to segmentation according to the above principles, until all the sub table does not exceed 1, the records to be sorted sequence becomes an ordered list. Assumed to be dividing sequence is r[left], r[left+1],... R[right], the realization process of division, can set two pointer in I and j, their initial values are left and right. The first reference recording r[left] to the variable x, r[left], r[i] is equivalent to the empty cell, and then repeated the following two scanning process, until I and j: The j scan from right to left, until r[j].key x.key, r[j] moved to the empty cell r[i], then r[j] is equivalent to the empty cell. The I scanned from left to right, until the r[i].key x.key, r[i] moved to the empty cell r[j], then r[i] is equivalent to the empty cell. When I and j met r