/ Check-in [25033ee9]

Overview
Comment: Corrections to the IN-operator notes. Tarball | ZIP archive | SQL archive family | ancestors | descendants | both | files | file ages | folders 25033ee94538289ba7e0147da30a18300047123f drh 2016-08-25 14:23:59
Context
 2016-08-25 15:46 Improvements to IN operator code generator comments. Avoid unnecessary Copy operations on the LHS of the IN operator. check-in: b6344298 user: drh tags: rowvalue 14:23 Corrections to the IN-operator notes. check-in: 25033ee9 user: drh tags: rowvalue 14:00 Add notes on the implementation of the IN operator. check-in: d256b2ca user: drh tags: rowvalue
Changes

Changes to src/in-operator.md.

```    58     58     4.  Return FALSE
59     59
60     60   ## Optimized Algorithm
61     61
62     62   The following procedure computes the same answer as the simple full-scan
63     63   algorithm, though it does so with less work in the common case.  This
64     64   is the algorithm that is implemented in SQLite.  The steps must occur
65         -in the order specified.  Except for the INDEX_NOOP optimization of step 1,
66         -none of the steps can be skipped.
65  +in the order specified.  Steps 1 and 3 are optional.  All other steps
66  +are required for correctness.
67     67
68     68     1.  If the RHS is a constant list of length 1 or 2, then rewrite the
69     69         IN operator as a simple expression.  Implement
70     70
71     71               x IN (y1,y2)
72     72
73     73         as if it were
................................................................................
76     76
77     77         This is the INDEX_NOOP optimization and is only undertaken if the
78     78         IN operator is used for membership testing.  If the IN operator is
79     79         driving a loop, then skip this step entirely.
80     80
81     81     2.  If the RHS is empty, return FALSE.
82     82
83         -  3.  If the LHS is a total-NULL or if the RHS contains a total-NULL,
84         -      then return NULL.
83  +  3.  If the LHS is a total-NULL, then return NULL.
85     84
86     85     4.  If the LHS is non-NULL, then use the LHS as a probe in a binary
87     86         search of the RHS
88     87
89         -      <ol type='a'>
90         -      <li> If the binary search finds an exact match, return TRUE
88  +      4-A.  If the binary search finds an exact match, return TRUE
91     89
92         -      <li> If the RHS is known to be not-null, return FALSE
93         -      </ol>
90  +      4-B.  If the RHS is known to be not-null, return FALSE
94     91
95     92     5.  At this point, it is known that the result cannot be TRUE.  All
96     93         that remains is to distinguish between NULL and FALSE.
97     94         If a NOT-TRUE result is acceptable, then return NOT-TRUE now.
98     95
99     96     6.  For each row in the RHS, compare that row against the LHS and
100     97         if the result is NULL, immediately return NULL.  This step is
101     98         essentially the "Simple Full-scan Algorithm" above with the
102     99         tests for TRUE removed, since we know that the result cannot be
103    100         TRUE at this point.
104    101
105    102     7.  Return FALSE.

```