Documentation Source Text

<dd><p>
^(Let X be U-35.  If the payload size P is less than or equal to X then
the entire payload is stored on the b-tree leaf page.)^
^(Let M be ((U-12)*32/255)-23 and let K be M+((P-M)%(U-4)).
If P is greater than X
then the number of bytes stored on the table b-tree leaf page is K
if K is less or equal to X or M otherwise.)^
^(The number of bytes stored on the leaf page is never less than M.)^
</p></dd>

<dt><p>Table B-Tree Interior Cell:</dt>
<dd><p>
Interior pages of table b-trees have no payload and so there is never
any payload to spill.
</p></dd>

<dt><p>Index B-Tree Leaf Or Interior Cell:</dt>
<dd><p>
^(Let X be ((U-12)*64/255)-23).  If the payload size P is less than
or equal to X then the entire payload is stored on the b-tree page.)^
^(Let M be ((U-12)*32/255)-23 and let K be M+((P-M)%(U-4)).
If P is greater than X then the number
of bytes stored on the index b-tree page is K if K is less than or
equal to X or M otherwise.)^
^(The number of bytes stored on the index page is never less than M.)^
</p></dd>
</dl>

<p>Here is an alternative description of the same computation:

<ul>
<li>X is U-35 for table btree leaf pages or

^Value is a BLOB that is (N-12)/2 bytes in length.
<tr><td valign=top align=center>N&#x2265;13 and odd
    <td valign=top align=center>(N-13)/2<td align=left>
^Value is a string in the [text encoding] and (N-13)/2 bytes in length.
^The nul terminator is not stored.
</table></center>

<p>The header size varint
and serial type varints will usually consist of a single byte.  The
serial type varints for large strings and BLOBs might extend to two or three
byte varints, but that is the exception rather than the rule. 
The varint format is very efficient at coding the record header.</p>

<p>^The values for each column in the record immediately follow the header.
^(For serial types 0, 8, 9, 12, and 13, the value is zero bytes in
length.  If all columns are of these types then the body section of the
record is empty.)^</p>

<p>^A record might have fewer values than the number of columns in the
corresponding table.  This can happen, for example, after an
[ALTER TABLE|ALTER TABLE ... ADD COLUMN] SQL statement has increased
the number of columns in the table schema without modifying preexisting rows
in the table.
^Missing values at the end of the record are filled in using the
[default value] for the corresponding columns defined in the table schema.
</p>


<h2>Record Sort Order</h2>

<p>The order of keys in an index b-tree is determined by the sort order of
the records that the keys represent.  Record comparison progresses column
by column.  Columns of a record are examined from left to right.  The
first pair of columns that are not equal determines the relative order

    <td> ^(The built-in BINARY collation compares strings byte by byte
        using the memcmp() function
        from the standard C library.)^
<tr><td valign=top>NOCASE
    <td> ^(The NOCASE collation is like BINARY except that uppercase
        ASCII characters ('A' through 'Z')
        are folded into their lowercase equivalents prior to running the
        comparison.  Only ASCII characters are case-folded.)^
        ^(NOCASE
        does not implement a general purpose unicode caseless comparison.)^
<tr><td valign=top>RTRIM
    <td> ^(RTRIM is like BINARY except that extra spaces at the end of either
         string do not change the result.  In other words, strings will
         compare equal to one another as long as they
         differ only in the number of spaces at the end.)^