给定一个索引项 p 和一个查询值 q，此函数确定该索引项是否与查询“一致”；也就是说，谓词“indexed_column indexable_operator q”对于该索引项所代表的任何行是否可能为真？对于叶子索引项，这相当于测试可索引条件，而对于内部树节点，这决定了是否有必要扫描该树节点所代表的索引子树。当结果为 true 时，还必须返回一个 recheck 标志。这表示谓词是确定为真还是仅可能为真。如果 recheck = false，则索引已精确测试了谓词条件；而如果 recheck = true，则该行只是一个候选匹配。在这种情况下，系统将自动对实际行值评估 indexable_operator，以查看它是否真的是一个匹配项。这个约定允许 GiSTGiST支持无损和有损索引结构。

该SQL函数的 SQL 声明必须如下所示：CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal) RETURNS bool

CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal)
RETURNS bool
AS 'MODULE_PATHNAME'
LANGUAGE C STRICT;

然后，C 模块中匹配的代码可以遵循这个骨架：PG_FUNCTION_INFO_V1(my_consistent); Datum my_consistent(PG_FUNCTION_ARGS) { GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); data_type *query = (data_type *) PG_GETARG_POINTER(1); StrategyNumber strategy = (StrategyNumber) PG_GETARG_UINT16(2); /* Oid subtype = PG_GETARG_OID(3); */ /* not used */ bool *recheck = (bool *) PG_GETARG_POINTER(4); data_type *key = DatumGetDataType(entry->key); bool retval; /* * set *recheck to false if the operator is exact for this data type, * true if it might be inexact. */ *recheck = true; /* * determine return value... */ retval = true; PG_RETURN_BOOL(retval); }

PG_FUNCTION_INFO_V1(my_consistent);

Datum
my_consistent(PG_FUNCTION_ARGS)
{
    GISTENTRY  *entry = (GISTENTRY *) PG_GETARG_POINTER(0);
    data_type  *query = PG_GETARG_DATA_TYPE_P(1);
    StrategyNumber strategy = (StrategyNumber) PG_GETARG_UINT16(2);
    /* Oid subtype = PG_GETARG_OID(3); */
    bool       *recheck = (bool *) PG_GETARG_POINTER(4);
    data_type  *key = DatumGetDataType(entry->key);
    bool        retval;

    /*
     * determine return value as a function of strategy, key and query.
     *
     * Use GIST_LEAF(entry) to know where you're called in the index tree,
     * which comes handy when supporting the = operator for example (you could
     * check for non empty union() in non-leaf nodes and equality in leaf
     * nodes).
     */

    *recheck = true;        /* or false if check is exact */

    PG_RETURN_BOOL(retval);
}

在这里，key 是索引中的一个元素，query 是在索引中查找的值。StrategyNumber 参数指示正在应用操作符类中的哪个操作符——它匹配 CREATE OPERATOR CLASS 命令中的一个操作符编号。

根据你在类中包含的操作符，query 的数据类型可能会随操作符而变化，因为它将是操作符右侧的任何类型，这可能与左侧出现的索引数据类型不同。（上面的代码骨架假设只有一种类型是可能的；如果不是，获取 query 参数值将不得不依赖于操作符。）建议 consistent 函数的 SQL 声明对 query 参数使用操作符类的索引数据类型，即使实际类型可能因操作符而异。

此方法合并树中的信息。给定一组条目，此函数会生成一个新的索引条目，该条目代表所有给定的条目。CREATE OR REPLACE FUNCTION my_union(internal, internal) RETURNS storage_typePG_FUNCTION_INFO_V1(my_union); Datum my_union(PG_FUNCTION_ARGS) { GistEntryVector *entryvec = (GistEntryVector *) PG_GETARG_POINTER(0); /* int *v_siz = (int *) PG_GETARG_POINTER(1); */ storage_type *out, *tmp; int i; out = palloc(sizeof(storage_type)); memcpy(out, DatumGetPointer(entryvec->vector[0].key), sizeof(storage_type)); for(i=1; i < entryvec->n; i++) { tmp = DatumGetPointer(entryvec->vector[i].key); /* a_union_of(out, tmp) */ memcpy(out, tmp, sizeof(storage_type)); } /* *v_siz = sizeof(storage_type); */ PG_RETURN_POINTER(out); }

该SQL函数的 SQL 声明必须如下所示：CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal) RETURNS bool

CREATE OR REPLACE FUNCTION my_union(internal, internal)
RETURNS storage_type
AS 'MODULE_PATHNAME'
LANGUAGE C STRICT;

然后，C 模块中匹配的代码可以遵循这个骨架：PG_FUNCTION_INFO_V1(my_consistent); Datum my_consistent(PG_FUNCTION_ARGS) { GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); data_type *query = (data_type *) PG_GETARG_POINTER(1); StrategyNumber strategy = (StrategyNumber) PG_GETARG_UINT16(2); /* Oid subtype = PG_GETARG_OID(3); */ /* not used */ bool *recheck = (bool *) PG_GETARG_POINTER(4); data_type *key = DatumGetDataType(entry->key); bool retval; /* * set *recheck to false if the operator is exact for this data type, * true if it might be inexact. */ *recheck = true; /* * determine return value... */ retval = true; PG_RETURN_BOOL(retval); }

PG_FUNCTION_INFO_V1(my_union);

Datum
my_union(PG_FUNCTION_ARGS)
{
    GistEntryVector *entryvec = (GistEntryVector *) PG_GETARG_POINTER(0);
    GISTENTRY  *ent = entryvec->vector;
    data_type  *out,
               *tmp,
               *old;
    int         numranges,
                i = 0;

    numranges = entryvec->n;
    tmp = DatumGetDataType(ent[0].key);
    out = tmp;

    if (numranges == 1)
    {
        out = data_type_deep_copy(tmp);

        PG_RETURN_DATA_TYPE_P(out);
    }

    for (i = 1; i < numranges; i++)
    {
        old = out;
        tmp = DatumGetDataType(ent[i].key);
        out = my_union_implementation(out, tmp);
    }

    PG_RETURN_DATA_TYPE_P(out);
}

如你所见，在这个骨架中，我们处理的是一种数据类型，其中 union(X, Y, Z) = union(union(X, Y), Z)。通过在这个 GiSTGiST支持方法中实现正确的联合算法，可以很容易地支持不符合这种情况的数据类型。

union 函数的结果必须是索引的存储类型的值，无论该类型是什么（它可能与索引列的类型不同，也可能相同）。union 函数应该返回一个指向新 palloc()ed 内存的指针。你不能仅仅按原样返回输入值，即使没有类型变化。

如上所示，union 函数的第一个 internal 参数实际上是一个 GistEntryVector 指针。第二个参数是一个指向整数变量的指针，可以忽略。（以前要求 union 函数将其结果值的大小存储在该变量中，但现在不再需要了。）

将数据项转换为适合在索引页中物理存储的格式。如果省略 compress 方法，数据项将不加修改地存储在索引中。CREATE OR REPLACE FUNCTION my_compress(internal) RETURNS internalPG_FUNCTION_INFO_V1(my_compress); Datum my_compress(PG_FUNCTION_ARGS) { GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); if (entry->leafkey) { /* leaf node */ GISTENTRY *retval; compressed_data_type *compressed_data = palloc(sizeof(compressed_data_type)); data_type *leaf_data = DatumGetPointer(entry->key); /* * convert leaf_data to compressed_data format... */ retval = palloc(sizeof(GISTENTRY)); gistentryinit(*retval, PointerGetDatum(compressed_data), entry->rel, entry->page, entry->offset, FALSE); PG_RETURN_POINTER(retval); } else { /* internal node */ PG_RETURN_POINTER(entry); } }

该SQL函数的 SQL 声明必须如下所示：CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal) RETURNS bool

CREATE OR REPLACE FUNCTION my_compress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C STRICT;

然后，C 模块中匹配的代码可以遵循这个骨架：PG_FUNCTION_INFO_V1(my_consistent); Datum my_consistent(PG_FUNCTION_ARGS) { GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); data_type *query = (data_type *) PG_GETARG_POINTER(1); StrategyNumber strategy = (StrategyNumber) PG_GETARG_UINT16(2); /* Oid subtype = PG_GETARG_OID(3); */ /* not used */ bool *recheck = (bool *) PG_GETARG_POINTER(4); data_type *key = DatumGetDataType(entry->key); bool retval; /* * set *recheck to false if the operator is exact for this data type, * true if it might be inexact. */ *recheck = true; /* * determine return value... */ retval = true; PG_RETURN_BOOL(retval); }

PG_FUNCTION_INFO_V1(my_compress);

Datum
my_compress(PG_FUNCTION_ARGS)
{
    GISTENTRY  *entry = (GISTENTRY *) PG_GETARG_POINTER(0);
    GISTENTRY  *retval;

    if (entry->leafkey)
    {
        /* replace entry->key with a compressed version */
        compressed_data_type *compressed_data = palloc(sizeof(compressed_data_type));

        /* fill *compressed_data from entry->key ... */

        retval = palloc(sizeof(GISTENTRY));
        gistentryinit(*retval, PointerGetDatum(compressed_data),
                      entry->rel, entry->page, entry->offset, FALSE);
    }
    else
    {
        /* typically we needn't do anything with non-leaf entries */
        retval = entry;
    }

    PG_RETURN_POINTER(retval);
}

当然，你必须将 compressed_data_type 调整为你正在转换的具体类型，以便压缩你的叶子节点。

将数据项的存储表示转换为可由操作符类中其他 GiST 方法操作的格式。如果省略 decompress 方法，则假定其他 GiST 方法可以直接处理存储的数据格式。（decompress 不一定是 compress 方法的逆过程；特别是，如果 compress 是有损的，那么 decompress 就不可能精确地重建原始数据。decompress 也不一定等同于 fetch，因为其他 GiST 方法可能不需要完全重建数据。）CREATE OR REPLACE FUNCTION my_decompress(internal) RETURNS internalPG_FUNCTION_INFO_V1(my_decompress); Datum my_decompress(PG_FUNCTION_ARGS) { PG_RETURN_POINTER(PG_GETARG_POINTER(0)); }

该SQL函数的 SQL 声明必须如下所示：CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal) RETURNS bool

CREATE OR REPLACE FUNCTION my_decompress(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C STRICT;

然后，C 模块中匹配的代码可以遵循这个骨架：PG_FUNCTION_INFO_V1(my_consistent); Datum my_consistent(PG_FUNCTION_ARGS) { GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); data_type *query = (data_type *) PG_GETARG_POINTER(1); StrategyNumber strategy = (StrategyNumber) PG_GETARG_UINT16(2); /* Oid subtype = PG_GETARG_OID(3); */ /* not used */ bool *recheck = (bool *) PG_GETARG_POINTER(4); data_type *key = DatumGetDataType(entry->key); bool retval; /* * set *recheck to false if the operator is exact for this data type, * true if it might be inexact. */ *recheck = true; /* * determine return value... */ retval = true; PG_RETURN_BOOL(retval); }

PG_FUNCTION_INFO_V1(my_decompress);

Datum
my_decompress(PG_FUNCTION_ARGS)
{
    PG_RETURN_POINTER(PG_GETARG_POINTER(0));
}

上面的骨架适用于不需要解压缩的情况。（但是，当然，完全省略该方法更容易，并且在这种情况下推荐这样做。）

返回一个值，表示将新条目插入树的特定分支的“成本”。项目将沿着树中 penalty 最小的路径插入。由 penalty 返回的值应为非负数。如果返回负值，它将被视为零。CREATE OR REPLACE FUNCTION my_penalty(internal, internal, internal) RETURNS internalPG_FUNCTION_INFO_V1(my_penalty); Datum my_penalty(PG_FUNCTION_ARGS) { GISTENTRY *origentry = (GISTENTRY *) PG_GETARG_POINTER(0); GISTENTRY *newentry = (GISTENTRY *) PG_GETARG_POINTER(1); float *penalty = (float *) PG_GETARG_POINTER(2); storage_type *orig = DatumGetPointer(origentry->key); storage_type *new = DatumGetPointer(newentry->key); /* * calculate penalty value, and store it in *penalty. */ *penalty = 0.0f; PG_RETURN_POINTER(penalty); }

该SQL函数的 SQL 声明必须如下所示：CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal) RETURNS bool

CREATE OR REPLACE FUNCTION my_penalty(internal, internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C STRICT;  -- in some cases penalty functions need not be strict

然后，C 模块中匹配的代码可以遵循这个骨架：PG_FUNCTION_INFO_V1(my_consistent); Datum my_consistent(PG_FUNCTION_ARGS) { GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); data_type *query = (data_type *) PG_GETARG_POINTER(1); StrategyNumber strategy = (StrategyNumber) PG_GETARG_UINT16(2); /* Oid subtype = PG_GETARG_OID(3); */ /* not used */ bool *recheck = (bool *) PG_GETARG_POINTER(4); data_type *key = DatumGetDataType(entry->key); bool retval; /* * set *recheck to false if the operator is exact for this data type, * true if it might be inexact. */ *recheck = true; /* * determine return value... */ retval = true; PG_RETURN_BOOL(retval); }

PG_FUNCTION_INFO_V1(my_penalty);

Datum
my_penalty(PG_FUNCTION_ARGS)
{
    GISTENTRY  *origentry = (GISTENTRY *) PG_GETARG_POINTER(0);
    GISTENTRY  *newentry = (GISTENTRY *) PG_GETARG_POINTER(1);
    float      *penalty = (float *) PG_GETARG_POINTER(2);
    data_type  *orig = DatumGetDataType(origentry->key);
    data_type  *new = DatumGetDataType(newentry->key);

    *penalty = my_penalty_implementation(orig, new);
    PG_RETURN_POINTER(penalty);
}

由于历史原因，penalty 函数不只是返回一个 float 结果；相反，它必须将值存储在由第三个参数指示的位置。返回值本身被忽略，尽管通常的做法是返回该参数的地址。

penalty 函数对索引的良好性能至关重要。它将在插入时用于确定在树中添加新条目时要遵循哪个分支。在查询时，索引越平衡，查找就越快。

picksplit 当需要进行索引页拆分时，此函数决定页面上的哪些条目将保留在旧页面上，哪些将移动到新页面上。CREATE OR REPLACE FUNCTION my_picksplit(internal, internal) RETURNS internalPG_FUNCTION_INFO_V1(my_picksplit); Datum my_picksplit(PG_FUNCTION_ARGS) { GistEntryVector *entryvec = (GistEntryVector *) PG_GETARG_POINTER(0); GIST_SPLITVEC *v = (GIST_SPLITVEC *) PG_GETARG_POINTER(1); int i; OffsetNumber maxoff; int nbytes; OffsetNumber *left, *right; data_type *tmp_union; data_type *left_union, *right_union; GISTENTRY *ent; maxoff = entryvec->n - 1; nbytes = (maxoff + 1) * sizeof(OffsetNumber); v->spl_left = (OffsetNumber *) palloc(nbytes); left = v->spl_left; v->spl_nleft = 0; v->spl_right = (OffsetNumber *) palloc(nbytes); right = v->spl_right; v->spl_nright = 0; left_union = NULL; right_union = NULL; for (i = FirstOffsetNumber; i <= maxoff; i = OffsetNumberNext(i)) { ent = &entryvec->vector[i]; tmp_union = DatumGetPointer(ent->key); /* * choose where to put the given entry, and update left_union and * right_union to include it. */ } v->spl_ldatum = PointerGetDatum(left_union); v->spl_rdatum = PointerGetDatum(right_union); PG_RETURN_POINTER(v); }

该SQL函数的 SQL 声明必须如下所示：CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal) RETURNS bool

CREATE OR REPLACE FUNCTION my_picksplit(internal, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C STRICT;

然后，C 模块中匹配的代码可以遵循这个骨架：PG_FUNCTION_INFO_V1(my_consistent); Datum my_consistent(PG_FUNCTION_ARGS) { GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); data_type *query = (data_type *) PG_GETARG_POINTER(1); StrategyNumber strategy = (StrategyNumber) PG_GETARG_UINT16(2); /* Oid subtype = PG_GETARG_OID(3); */ /* not used */ bool *recheck = (bool *) PG_GETARG_POINTER(4); data_type *key = DatumGetDataType(entry->key); bool retval; /* * set *recheck to false if the operator is exact for this data type, * true if it might be inexact. */ *recheck = true; /* * determine return value... */ retval = true; PG_RETURN_BOOL(retval); }

PG_FUNCTION_INFO_V1(my_picksplit);

Datum
my_picksplit(PG_FUNCTION_ARGS)
{
    GistEntryVector *entryvec = (GistEntryVector *) PG_GETARG_POINTER(0);
    GIST_SPLITVEC *v = (GIST_SPLITVEC *) PG_GETARG_POINTER(1);
    OffsetNumber maxoff = entryvec->n - 1;
    GISTENTRY  *ent = entryvec->vector;
    int         i,
                nbytes;
    OffsetNumber *left,
               *right;
    data_type  *tmp_union;
    data_type  *unionL;
    data_type  *unionR;
    GISTENTRY **raw_entryvec;

    maxoff = entryvec->n - 1;
    nbytes = (maxoff + 1) * sizeof(OffsetNumber);

    v->spl_left = (OffsetNumber *) palloc(nbytes);
    left = v->spl_left;
    v->spl_nleft = 0;

    v->spl_right = (OffsetNumber *) palloc(nbytes);
    right = v->spl_right;
    v->spl_nright = 0;

    unionL = NULL;
    unionR = NULL;

    /* Initialize the raw entry vector. */
    raw_entryvec = (GISTENTRY **) malloc(entryvec->n * sizeof(void *));
    for (i = FirstOffsetNumber; i <= maxoff; i = OffsetNumberNext(i))
        raw_entryvec[i] = &(entryvec->vector[i]);

    for (i = FirstOffsetNumber; i <= maxoff; i = OffsetNumberNext(i))
    {
        int         real_index = raw_entryvec[i] - entryvec->vector;

        tmp_union = DatumGetDataType(entryvec->vector[real_index].key);
        Assert(tmp_union != NULL);

        /*
         * Choose where to put the index entries and update unionL and unionR
         * accordingly. Append the entries to either v->spl_left or
         * v->spl_right, and care about the counters.
         */

        if (my_choice_is_left(unionL, curl, unionR, curr))
        {
            if (unionL == NULL)
                unionL = tmp_union;
            else
                unionL = my_union_implementation(unionL, tmp_union);

            *left = real_index;
            ++left;
            ++(v->spl_nleft);
        }
        else
        {
            /*
             * Same on the right
             */
        }
    }

    v->spl_ldatum = DataTypeGetDatum(unionL);
    v->spl_rdatum = DataTypeGetDatum(unionR);
    PG_RETURN_POINTER(v);
}

请注意，picksplit 函数的结果是通过修改传入的 v 结构体来传递的。返回值本身被忽略，尽管通常的做法是返回 v 的地址。

与 penalty 类似，picksplit 函数对索引的良好性能至关重要。设计合适的 penalty 和 picksplit 实现是实现高性能 GiSTGiST索引的挑战所在。

如果两个索引条目相同则返回 true，否则返回 false。（“索引条目”是索引存储类型的值，不一定是原始索引列的类型。）CREATE OR REPLACE FUNCTION my_same(storage_type, storage_type, internal) RETURNS internalPG_FUNCTION_INFO_V1(my_same); Datum my_same(PG_FUNCTION_ARGS) { storage_type *b1 = (storage_type *) PG_GETARG_POINTER(0); storage_type *b2 = (storage_type *) PG_GETARG_POINTER(1); bool *result = (bool *) PG_GETARG_POINTER(2); /* * compare b1 and b2, and set *result */ *result = false; PG_RETURN_POINTER(result); }

该SQL函数的 SQL 声明必须如下所示：CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal) RETURNS bool

CREATE OR REPLACE FUNCTION my_same(storage_type, storage_type, internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C STRICT;

然后，C 模块中匹配的代码可以遵循这个骨架：PG_FUNCTION_INFO_V1(my_consistent); Datum my_consistent(PG_FUNCTION_ARGS) { GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); data_type *query = (data_type *) PG_GETARG_POINTER(1); StrategyNumber strategy = (StrategyNumber) PG_GETARG_UINT16(2); /* Oid subtype = PG_GETARG_OID(3); */ /* not used */ bool *recheck = (bool *) PG_GETARG_POINTER(4); data_type *key = DatumGetDataType(entry->key); bool retval; /* * set *recheck to false if the operator is exact for this data type, * true if it might be inexact. */ *recheck = true; /* * determine return value... */ retval = true; PG_RETURN_BOOL(retval); }

PG_FUNCTION_INFO_V1(my_same);

Datum
my_same(PG_FUNCTION_ARGS)
{
    prefix_range *v1 = PG_GETARG_PREFIX_RANGE_P(0);
    prefix_range *v2 = PG_GETARG_PREFIX_RANGE_P(1);
    bool       *result = (bool *) PG_GETARG_POINTER(2);

    *result = my_eq(v1, v2);
    PG_RETURN_POINTER(result);
}

由于历史原因，same 函数不只是返回一个布尔结果；相反，它必须将标志存储在由第三个参数指示的位置。返回值本身被忽略，尽管通常的做法是返回该参数的地址。

给定一个索引项 p 和一个查询值 q，此函数确定该索引项与查询值的“距离”。如果操作符类包含任何排序操作符，则必须提供此函数。使用排序操作符的查询将通过首先返回“距离”值最小的索引项来实现，因此结果必须与操作符的语义一致。对于叶子索引项，结果仅表示到该索引项的距离；对于内部树节点，结果必须是任何子项可能具有的最小距离。CREATE OR REPLACE FUNCTION my_distance(internal, data_type, smallint, oid, internal) RETURNS float8

该SQL函数的 SQL 声明必须如下所示：CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal) RETURNS bool

CREATE OR REPLACE FUNCTION my_distance(internal, data_type, smallint, oid, internal)
RETURNS float8
AS 'MODULE_PATHNAME'
LANGUAGE C STRICT;

然后，C 模块中匹配的代码可以遵循这个骨架：PG_FUNCTION_INFO_V1(my_consistent); Datum my_consistent(PG_FUNCTION_ARGS) { GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); data_type *query = (data_type *) PG_GETARG_POINTER(1); StrategyNumber strategy = (StrategyNumber) PG_GETARG_UINT16(2); /* Oid subtype = PG_GETARG_OID(3); */ /* not used */ bool *recheck = (bool *) PG_GETARG_POINTER(4); data_type *key = DatumGetDataType(entry->key); bool retval; /* * set *recheck to false if the operator is exact for this data type, * true if it might be inexact. */ *recheck = true; /* * determine return value... */ retval = true; PG_RETURN_BOOL(retval); }

PG_FUNCTION_INFO_V1(my_distance);

Datum
my_distance(PG_FUNCTION_ARGS)
{
    GISTENTRY  *entry = (GISTENTRY *) PG_GETARG_POINTER(0);
    data_type  *query = PG_GETARG_DATA_TYPE_P(1);
    StrategyNumber strategy = (StrategyNumber) PG_GETARG_UINT16(2);
    /* Oid subtype = PG_GETARG_OID(3); */
    /* bool *recheck = (bool *) PG_GETARG_POINTER(4); */
    data_type  *key = DatumGetDataType(entry->key);
    double      retval;

    /*
     * determine return value as a function of strategy, key and query.
     */

    PG_RETURN_FLOAT8(retval);
}

distance 函数的参数与 consistent 函数的参数相同。

在确定距离时允许某些近似，只要结果永远不大于条目的实际距离。因此，例如，在几何应用中，到边界框的距离通常是足够的。对于内部树节点，返回的距离不得大于到任何子节点的距离。如果返回的距离不精确，函数必须将 *recheck 设置为 true。（对于内部树节点，这不是必需的；对于它们，计算总是被假定为不精确的。）在这种情况下，执行器将在从堆中获取元组后计算准确的距离，并在必要时对元组重新排序。

如果距离函数对任何叶子节点返回 *recheck = true，则原始排序操作符的返回类型必须是 float8 或 float4，并且距离函数的结果值必须与原始排序操作符的结果值可比，因为执行器将同时使用距离函数结果和重新计算的排序操作符结果进行排序。否则，距离函数的结果值可以是任何有限的 float8 值，只要结果值的相对顺序与排序操作符返回的顺序匹配。（无穷大和负无穷大在内部用于处理诸如 null 的情况，因此不建议 distance 函数返回这些值。）

为仅索引扫描将数据项的压缩索引表示转换为原始数据类型。返回的数据必须是原始索引值的精确、无损副本。

该SQL函数的 SQL 声明必须如下所示：CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal) RETURNS bool

CREATE OR REPLACE FUNCTION my_fetch(internal)
RETURNS internal
AS 'MODULE_PATHNAME'
LANGUAGE C STRICT;

参数是指向 GISTENTRY 结构体的指针。在进入时，其 key 字段包含一个非 NULL 的压缩形式的叶子数据。返回值是另一个 GISTENTRY 结构体，其 key 字段包含原始、未压缩形式的相同数据。如果操作符类的压缩函数对叶子条目不做任何操作，则 fetch 方法可以按原样返回参数。或者，如果操作符类没有压缩函数，fetch 方法也可以省略，因为它必然是无操作的。CREATE OR REPLACE FUNCTION my_fetch(internal) RETURNS internal

然后，C 模块中匹配的代码可以遵循这个骨架：PG_FUNCTION_INFO_V1(my_fetch); Datum my_fetch(PG_FUNCTION_ARGS) { GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); if (entry->leafkey) { GISTENTRY *retval; data_type *uncompressed_data = palloc(sizeof(data_type)); compressed_data_type *leaf_data = DatumGetPointer(entry->key); /* * convert leaf_data to uncompressed_data format... */ retval = palloc(sizeof(GISTENTRY)); gistentryinit(*retval, PointerGetDatum(uncompressed_data), entry->rel, entry->page, entry->offset, FALSE); PG_RETURN_POINTER(retval); } else { PG_RETURN_POINTER(entry); } }

PG_FUNCTION_INFO_V1(my_fetch);

Datum
my_fetch(PG_FUNCTION_ARGS)
{
    GISTENTRY  *entry = (GISTENTRY *) PG_GETARG_POINTER(0);
    input_data_type *in = DatumGetPointer(entry->key);
    fetched_data_type *fetched_data;
    GISTENTRY  *retval;

    retval = palloc(sizeof(GISTENTRY));
    fetched_data = palloc(sizeof(fetched_data_type));

    /*
     * Convert 'fetched_data' into the a Datum of the original datatype.
     */

    /* fill *retval from fetched_data. */
    gistentryinit(*retval, PointerGetDatum(converted_datum),
                  entry->rel, entry->page, entry->offset, FALSE);

    PG_RETURN_POINTER(retval);
}

如果压缩方法对叶子条目是有损的，则操作符类不能支持仅索引扫描，并且不得定义 fetch 函数。

options 允许定义用户可见的参数，以控制操作符类的行为。CREATE OR REPLACE FUNCTION my_options(internal) RETURNS void

该SQL函数的 SQL 声明必须如下所示：CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal) RETURNS bool

CREATE OR REPLACE FUNCTION my_options(internal)
RETURNS void
AS 'MODULE_PATHNAME'
LANGUAGE C STRICT;

该函数接收一个指向 local_relopts 结构的指针，该结构需要填充一组操作符类特定的选项。可以使用 PG_HAS_OPCLASS_OPTIONS() 和 PG_GET_OPCLASS_OPTIONS() 宏从其他支持函数访问这些选项。

下面给出了一个 my_options() 和其他支持函数使用参数的实现示例：typedef struct { int32 vl_len_; /* varlena header */ int siglen; } MyOptions; ... PG_FUNCTION_INFO_V1(my_options); Datum my_options(PG_FUNCTION_ARGS) { local_relopts *relopts = (local_relopts *) PG_GETARG_POINTER(0); int siglen = 128; MyOptions *options; /* Get integer parameter siglen from user. */ get_reloptions_int("siglen", &siglen, relopts, 1, 2048, 10); options = (MyOptions *) palloc(sizeof(MyOptions)); SET_VARSIZE(options, sizeof(MyOptions)); options->siglen = siglen; PG_RETURN_POINTER(options); } ... /* In other support functions */ MyOptions *options = (MyOptions *) PG_GET_OPCLASS_OPTIONS(fcinfo->flinfo);

typedef enum MyEnumType
{
    MY_ENUM_ON,
    MY_ENUM_OFF,
    MY_ENUM_AUTO
} MyEnumType;

typedef struct
{
    int32   vl_len_;    /* varlena header (do not touch directly!) */
    int     int_param;  /* integer parameter */
    double  real_param; /* real parameter */
    MyEnumType enum_param; /* enum parameter */
    int     str_param;  /* string parameter */
} MyOptionsStruct;

/* String representation of enum values */
static relopt_enum_elt_def myEnumValues[] =
{
    {"on", MY_ENUM_ON},
    {"off", MY_ENUM_OFF},
    {"auto", MY_ENUM_AUTO},
    {(const char *) NULL}   /* list terminator */
};

static char *str_param_default = "default";

/*
 * Sample validator: checks that string is not longer than 8 bytes.
 */
static void
validate_my_string_relopt(const char *value)
{
    if (strlen(value) > 8)
        ereport(ERROR,
                (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
                 errmsg("str_param must be at most 8 bytes")));
}

/*
 * Sample filler: switches characters to lower case.
 */
static Size
fill_my_string_relopt(const char *value, void *ptr)
{
    char   *tmp = str_tolower(value, strlen(value), DEFAULT_COLLATION_OID);
    int     len = strlen(tmp);

    if (ptr)
        strcpy(ptr, tmp);

    pfree(tmp);
    return len + 1;
}

PG_FUNCTION_INFO_V1(my_options);

Datum
my_options(PG_FUNCTION_ARGS)
{
    local_relopts *relopts = (local_relopts *) PG_GETARG_POINTER(0);

    init_local_reloptions(relopts, sizeof(MyOptionsStruct));
    add_local_int_reloption(relopts, "int_param", "integer parameter",
                            100, 0, 1000000,
                            offsetof(MyOptionsStruct, int_param));
    add_local_real_reloption(relopts, "real_param", "real parameter",
                             1.0, 0.0, 1000000.0,
                             offsetof(MyOptionsStruct, real_param));
    add_local_enum_reloption(relopts, "enum_param", "enum parameter",
                             myEnumValues, MY_ENUM_ON,
                             "Valid values are: \"on\", \"off\" and \"auto\".",
                             offsetof(MyOptionsStruct, enum_param));
    add_local_string_reloption(relopts, "str_param", "string parameter",
                               str_param_default,
                               &validate_my_string_relopt,
                               &fill_my_string_relopt,
                               offsetof(MyOptionsStruct, str_param));

    PG_RETURN_VOID();
}

PG_FUNCTION_INFO_V1(my_compress);

Datum
my_compress(PG_FUNCTION_ARGS)
{
    int     int_param = 100;
    double  real_param = 1.0;
    MyEnumType enum_param = MY_ENUM_ON;
    char   *str_param = str_param_default;

    /*
     * Normally, when opclass contains 'options' method, then options are always
     * passed to support functions.  However, if you add 'options' method to
     * existing opclass, previously defined indexes have no options, so the
     * check is required.
     */
    if (PG_HAS_OPCLASS_OPTIONS())
    {
        MyOptionsStruct *options = (MyOptionsStruct *) PG_GET_OPCLASS_OPTIONS();

        int_param = options->int_param;
        real_param = options->real_param;
        enum_param = options->enum_param;
        str_param = GET_STRING_RELOPTION(options, str_param);
    }

    /* the rest implementation of support function */
}

由于 GiST 中键的表示是灵活的，GiST它可能依赖于用户指定的参数。例如，可以指定键签名的长度。参见 gtsvector_options() 的示例。

sortsupport 返回一个比较器函数，以一种保留局部性的方式对数据进行排序。它被 CREATE INDEX 和 REINDEX 命令使用。创建的索引的质量取决于比较器函数确定的排序顺序保留输入局部性的程度。CREATE OR REPLACE FUNCTION my_sortsupport(internal) RETURNS internal

sortsupport 方法是可选的。如果没有提供，CREATE INDEX 将通过使用 penalty 和 picksplit 函数将每个元组插入到树中来构建索引，这要慢得多。

该SQL函数的 SQL 声明必须如下所示：CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal) RETURNS bool

CREATE OR REPLACE FUNCTION my_sortsupport(internal)
RETURNS void
AS 'MODULE_PATHNAME'
LANGUAGE C STRICT;

参数是指向 SortSupport 结构体的指针。函数至少必须填充其比较器字段。比较器接受三个参数：要比较的两个 Datum，以及一个指向 SortSupport 结构体的指针。Datum 是存储在索引中的两个索引值；也就是说，是 compress 方法返回的格式。完整的 API 定义在 src/include/utils/sortsupport.h 中。

然后，C 模块中匹配的代码可以遵循这个骨架：PG_FUNCTION_INFO_V1(my_fetch); Datum my_fetch(PG_FUNCTION_ARGS) { GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); if (entry->leafkey) { GISTENTRY *retval; data_type *uncompressed_data = palloc(sizeof(data_type)); compressed_data_type *leaf_data = DatumGetPointer(entry->key); /* * convert leaf_data to uncompressed_data format... */ retval = palloc(sizeof(GISTENTRY)); gistentryinit(*retval, PointerGetDatum(uncompressed_data), entry->rel, entry->page, entry->offset, FALSE); PG_RETURN_POINTER(retval); } else { PG_RETURN_POINTER(entry); } }

PG_FUNCTION_INFO_V1(my_sortsupport);

static int
my_fastcmp(Datum x, Datum y, SortSupport ssup)
{
  /* establish order between x and y by computing some sorting value z */

  int z1 = ComputeSpatialCode(x);
  int z2 = ComputeSpatialCode(y);

  return z1 == z2 ? 0 : z1 > z2 ? 1 : -1;
}

Datum
my_sortsupport(PG_FUNCTION_ARGS)
{
  SortSupport ssup = (SortSupport) PG_GETARG_POINTER(0);

  ssup->comparator = my_fastcmp;
  PG_RETURN_VOID();
}

给定一个来自 src/include/nodes/primnodes.h 的 CompareType 值，返回此操作符类用于匹配功能的策略号。如果操作符类没有匹配的策略，函数应返回 InvalidStrategy。CREATE OR REPLACE FUNCTION my_translate_cmptype(int) RETURNS smallint

这用于时态索引约束（即 PRIMARY KEY 和 UNIQUE）。如果操作符类提供了此函数并且它为 COMPARE_EQ 返回结果，则它可以在索引约束的非 WITHOUT OVERLAPS 部分中使用。

此支持函数对应于索引访问方法回调函数 amtranslatecmptype（参见 第 63.2 节）。GiST 索引的 amtranslatecmptype 回调函数仅调用相应操作符族的 translate_cmptype 支持函数，因为 GiST 索引访问方法本身没有固定的策略号。

该SQL函数的 SQL 声明必须如下所示：CREATE OR REPLACE FUNCTION my_consistent(internal, data_type, smallint, oid, internal) RETURNS bool

CREATE OR REPLACE FUNCTION my_translate_cmptype(integer)
RETURNS smallint
AS 'MODULE_PATHNAME'
LANGUAGE C STRICT;

并且操作符族注册必须如下所示：OPERATOR 1 =(any, any), ... , FUNCTION 12 my_translate_cmptype (int4)

ALTER OPERATOR FAMILY my_opfamily USING gist ADD
    FUNCTION 12 ("any", "any") my_translate_cmptype(int);

然后，C 模块中匹配的代码可以遵循这个骨架：PG_FUNCTION_INFO_V1(my_fetch); Datum my_fetch(PG_FUNCTION_ARGS) { GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); if (entry->leafkey) { GISTENTRY *retval; data_type *uncompressed_data = palloc(sizeof(data_type)); compressed_data_type *leaf_data = DatumGetPointer(entry->key); /* * convert leaf_data to uncompressed_data format... */ retval = palloc(sizeof(GISTENTRY)); gistentryinit(*retval, PointerGetDatum(uncompressed_data), entry->rel, entry->page, entry->offset, FALSE); PG_RETURN_POINTER(retval); } else { PG_RETURN_POINTER(entry); } }

PG_FUNCTION_INFO_V1(my_translate_cmptype);

Datum
my_translate_cmptype(PG_FUNCTION_ARGS)
{
    CompareType cmptype = PG_GETARG_INT32(0);
    StrategyNumber ret = InvalidStrategy;

    switch (cmptype)
    {
        case COMPARE_EQ:
            ret = BTEqualStrategyNumber;
    }

    PG_RETURN_UINT16(ret);
}

PostgreSQL 提供了一个翻译函数：gist_translate_cmptype_common 适用于使用 RT*StrategyNumber 常量的操作符类。btree_gist 扩展定义了第二个翻译函数，gist_translate_cmptype_btree，适用于使用 BT*StrategyNumber 常量的操作符类。

btree_gist 为多种数据类型提供 B 树等效功能

为多维立方体进行索引

hstore 用于存储（键，值）对的模块

intarray 用于 int4 值一维数组的 RD-Tree

ltree 为树状结构进行索引

pg_trgm 使用三元匹配进行文本相似度比较

seg 为“浮点数范围”进行索引