SQLite源码Btree的介绍

SQLite使用C开发，存储引擎部分使用的是B树。

对于存储的最小粒度页(page)来说，里面包含数据项和指向其他页的指针。
对于搜索磁盘上某一个key的值，需要读取树中O(log(M))的页。

** The basic idea is that each page of the file contains N database
** entries and N+1 pointers to subpages.
**
**   ----------------------------------------------------------------
**   |  Ptr(0) | Key(0) | Ptr(1) | Key(1) | ... | Key(N-1) | Ptr(N) |
**   ----------------------------------------------------------------

在树的实现上，一个单独的数据文件可以包含一颗或多颗独立的B树。
对于表中任何数据条目来说，对应的key和数据组成了payload，一定数据量的payload可以直接在页上面存放。
如果页上存放payload的空间不足，则会讲多余的数据存放到overflow页中。payload和前一项指针组成了Cell。
每个页有一个小的头，包含了Ptr(N)指针和其他信息，例如key和数据的大小。
页的种类，有btree、freelist、overflow和pointer-map。

每个btree页分成三部分，头部、cell指针数组和cell内容区。
第一个页还包含一个100字节的文件头。在磁盘上存储的格式如下。

**
**      |----------------|
**      | file header    |   100 bytes.  Page 1 only.
**      |----------------|
**      | page header    |   8 bytes for leaves.  12 bytes for interior nodes
**      |----------------|
**      | cell pointer   |   |  2 bytes per cell.  Sorted order.
**      | array          |   |  Grows downward
**      |                |   v
**      |----------------|
**      | unallocated    |
**      | space          |
**      |----------------|   ^  Grows upwards
**      | cell content   |   |  Arbitrary order interspersed with freeblocks.
**      | area           |   |  and free space fragments.
**      |----------------|

页头部如下

**   OFFSET   SIZE     DESCRIPTION
**      0       1      Flags. 1: intkey, 2: zerodata, 4: leafdata, 8: leaf
**      1       2      byte offset to the first freeblock
**      3       2      number of cells on this page
**      5       2      first byte of the cell content area
**      7       1      number of fragmented free bytes
**      8       4      Right child (the Ptr(N) value).  Omitted on leaves.

cell指针数组包含0个户多个2字节的指针，这些指针式cell内容区中对应的偏移量。
cell指针按照顺序存储，这样做的好处是，新的cell可以很容易添加而不会导致页的分裂。

cell的内容如下

**    SIZE    DESCRIPTION
**      4     Page number of the left child. Omitted if leaf flag is set.
**     var    Number of bytes of data. Omitted if the zerodata flag is set.
**     var    Number of bytes of key. Or the key itself if intkey flag is set.
**      *     Payload
**      4     First page of the overflow chain.  Omitted if no overflow

Freelist页有两类，trunk页和leaf页。

MemPage是单独数据库页加载到内存中的对象。对象中的信息来源于磁盘上页的内容。

struct MemPage {
  u8 isInit;           /* True if previously initialized. MUST BE FIRST! */
  u8 intKey;           /* True if table b-trees.  False for index b-trees */
  u8 intKeyLeaf;       /* True if the leaf of an intKey table */
  Pgno pgno;           /* Page number for this page */
  /* Only the first 8 bytes (above) are zeroed by pager.c when a new page
  ** is allocated. All fields that follow must be initialized before use */
  u8 leaf;             /* True if a leaf page */
  u8 hdrOffset;        /* 100 for page 1.  0 otherwise */
  u8 childPtrSize;     /* 0 if leaf==1.  4 if leaf==0 */
  u8 max1bytePayload;  /* min(maxLocal,127) */
  u8 nOverflow;        /* Number of overflow cell bodies in aCell[] */
  u16 maxLocal;        /* Copy of BtShared.maxLocal or BtShared.maxLeaf */
  u16 minLocal;        /* Copy of BtShared.minLocal or BtShared.minLeaf */
  u16 cellOffset;      /* Index in aData of first cell pointer */
  int nFree;           /* Number of free bytes on the page. -1 for unknown */
  u16 nCell;           /* Number of cells on this page, local and ovfl */
  u16 maskPage;        /* Mask for page offset */
  u16 aiOvfl[4];       /* Insert the i-th overflow cell before the aiOvfl-th
                       ** non-overflow cell */
  u8 *apOvfl[4];       /* Pointers to the body of overflow cells */
  BtShared *pBt;       /* Pointer to BtShared that this page is part of */
  u8 *aData;           /* Pointer to disk image of the page data */
  u8 *aDataEnd;        /* One byte past the end of the entire page - not just
                       ** the usable space, the entire page.  Used to prevent
                       ** corruption-induced buffer overflow. */
  u8 *aCellIdx;        /* The cell index area */
  u8 *aDataOfst;       /* Same as aData for leaves.  aData+4 for interior */
  DbPage *pDbPage;     /* Pager page handle */
  u16 (*xCellSize)(MemPage*,u8*);             /* cellSizePtr method */
  void (*xParseCell)(MemPage*,u8*,CellInfo*); /* btreeParseCell method */
};

Btree的句柄
当打开一个数据库文件时，数据库的连接包含了指向这个对象的指针。

struct Btree {
  sqlite3 *db;       /* The database connection holding this btree */
  BtShared *pBt;     /* Sharable content of this btree */
  u8 inTrans;        /* TRANS_NONE, TRANS_READ or TRANS_WRITE */
  u8 sharable;       /* True if we can share pBt with another db */
  u8 locked;         /* True if db currently has pBt locked */
  u8 hasIncrblobCur; /* True if there are one or more Incrblob cursors */
  int wantToLock;    /* Number of nested calls to sqlite3BtreeEnter() */
  int nBackup;       /* Number of backup operations reading this btree */
  u32 iBDataVersion; /* Combines with pBt->pPager->iDataVersion */
  Btree *pNext;      /* List of other sharable Btrees from the same db */
  Btree *pPrev;      /* Back pointer of the same list */
#ifdef SQLITE_DEBUG
  u64 nSeek;         /* Calls to sqlite3BtreeMovetoUnpacked() */
#endif
#ifndef SQLITE_OMIT_SHARED_CACHE
  BtLock lock;       /* Object used to lock page 1 */
#endif
};

BtShared数据文件实例
一个数据库在同一时间可以被两个或更多的数据库连接打开。当这些连接共享相同的数据库文件时，每个连接有它自己的私有Btree对象，每个Btree
会指向BtShared对象。

struct BtShared {
  Pager *pPager;        /* The page cache */
  sqlite3 *db;          /* Database connection currently using this Btree */
  BtCursor *pCursor;    /* A list of all open cursors */
  MemPage *pPage1;      /* First page of the database */
  u8 openFlags;         /* Flags to sqlite3BtreeOpen() */
#ifndef SQLITE_OMIT_AUTOVACUUM
  u8 autoVacuum;        /* True if auto-vacuum is enabled */
  u8 incrVacuum;        /* True if incr-vacuum is enabled */
  u8 bDoTruncate;       /* True to truncate db on commit */
#endif
  u8 inTransaction;     /* Transaction state */
  u8 max1bytePayload;   /* Maximum first byte of cell for a 1-byte payload */
  u8 nReserveWanted;    /* Desired number of extra bytes per page */
  u16 btsFlags;         /* Boolean parameters.  See BTS_* macros below */
  u16 maxLocal;         /* Maximum local payload in non-LEAFDATA tables */
  u16 minLocal;         /* Minimum local payload in non-LEAFDATA tables */
  u16 maxLeaf;          /* Maximum local payload in a LEAFDATA table */
  u16 minLeaf;          /* Minimum local payload in a LEAFDATA table */
  u32 pageSize;         /* Total number of bytes on a page */
  u32 usableSize;       /* Number of usable bytes on each page */
  int nTransaction;     /* Number of open transactions (read + write) */
  u32 nPage;            /* Number of pages in the database */
  void *pSchema;        /* Pointer to space allocated by sqlite3BtreeSchema() */
  void (*xFreeSchema)(void*);  /* Destructor for BtShared.pSchema */
  sqlite3_mutex *mutex; /* Non-recursive mutex required to access this object */
  Bitvec *pHasContent;  /* Set of pages moved to free-list this transaction */
#ifndef SQLITE_OMIT_SHARED_CACHE
  int nRef;             /* Number of references to this structure */
  BtShared *pNext;      /* Next on a list of sharable BtShared structs */
  BtLock *pLock;        /* List of locks held on this shared-btree struct */
  Btree *pWriter;       /* Btree with currently open write transaction */
#endif
  u8 *pTmpSpace;        /* Temp space sufficient to hold a single cell */
  int nPreformatSize;   /* Size of last cell written by TransferRow() */
};

BtCursor游标
游标时数据库文件中，指向特定b-tree的条目的指针，对应MemPage和MemPage.aCell[]中的索引。
数据库可以同时被多个数据连接共享，但是游标不可以共享。每个游标和BtCursor.pBtree.db关联。

struct BtCursor {
  u8 eState;                /* One of the CURSOR_XXX constants (see below) */
  u8 curFlags;              /* zero or more BTCF_* flags defined below */
  u8 curPagerFlags;         /* Flags to send to sqlite3PagerGet() */
  u8 hints;                 /* As configured by CursorSetHints() */
  int skipNext;    /* Prev() is noop if negative. Next() is noop if positive.
                   ** Error code if eState==CURSOR_FAULT */
  Btree *pBtree;            /* The Btree to which this cursor belongs */
  Pgno *aOverflow;          /* Cache of overflow page locations */
  void *pKey;               /* Saved key that was cursor last known position */
  /* All fields above are zeroed when the cursor is allocated.  See
  ** sqlite3BtreeCursorZero().  Fields that follow must be manually
  ** initialized. */
#define BTCURSOR_FIRST_UNINIT pBt   /* Name of first uninitialized field */
  BtShared *pBt;            /* The BtShared this cursor points to */
  BtCursor *pNext;          /* Forms a linked list of all cursors */
  CellInfo info;            /* A parse of the cell we are pointing at */
  i64 nKey;                 /* Size of pKey, or last integer key */
  Pgno pgnoRoot;            /* The root page of this tree */
  i8 iPage;                 /* Index of current page in apPage */
  u8 curIntKey;             /* Value of apPage[0]->intKey */
  u16 ix;                   /* Current index for apPage[iPage] */
  u16 aiIdx[BTCURSOR_MAX_DEPTH-1];     /* Current index in apPage[i] */
  struct KeyInfo *pKeyInfo;            /* Arg passed to comparison function */
  MemPage *pPage;                        /* Current page */
  MemPage *apPage[BTCURSOR_MAX_DEPTH-1]; /* Stack of parents of current page */
};