PostgreSQL 原始碼解讀（103）- 分割槽表#9（資料查詢路由#6-APPEND初始化和實現）
本節介紹了APPEND Plan Node節點的初始化和執行邏輯，主要的實現函式包括ExecInitAppend和ExecAppend。
一、資料結構
AppendState
用於Append Node執行的資料結構
/* ----------------
* AppendState information
* AppendState結構體
*
* nplans how many plans are in the array
* 陣列大小
* whichplan which plan is being executed (0 .. n-1), or a
* special negative value. See nodeAppend.c.
* 那個計劃將要被/正在被執行(0 .. n-1),或者一個特別的負數,詳見nodeAppend.c
* pruningstate details required to allow partitions to be
* eliminated from the scan, or NULL if not possible.
* 在掃描過程中允許忽略分割槽的細節資訊,入不可能則為NULL
* valid_subplans for runtime pruning, valid appendplans indexes to
* scan.
* 用於執行期裁剪分割槽,將要掃描的有效appendplans索引
* ----------------
*/
struct AppendState;
typedef struct AppendState AppendState;
struct ParallelAppendState;
typedef struct ParallelAppendState ParallelAppendState;
struct PartitionPruneState;
struct AppendState
{
//第一個欄位為NodeTag
PlanState ps; /* its first field is NodeTag */
//PlanStates陣列
PlanState **appendplans; /* array of PlanStates for my inputs */
//陣列大小
int as_nplans;
//那個計劃將要被/正在被執行
int as_whichplan;
//包含第一個部分計劃的appendplans陣列編號
int as_first_partial_plan; /* Index of 'appendplans' containing
* the first partial plan */
//並行執行的協調資訊
ParallelAppendState *as_pstate; /* parallel coordination info */
//協調資訊大小
Size pstate_len; /* size of parallel coordination info */
//分割槽裁剪資訊
struct PartitionPruneState *as_prune_state;
//有效的子計劃點陣圖集
Bitmapset *as_valid_subplans;
//選擇下個子計劃的實現函式
bool (*choose_next_subplan) (AppendState *);
};
二、原始碼解讀
ExecInitAppend
ExecInitAppend函式為開始append node的所有子關係掃描執行相關的初始化.
/* ----------------------------------------------------------------
* ExecInitAppend
*
* Begin all of the subscans of the append node.
* 開始append node的所有子關係掃描
*
* (This is potentially wasteful, since the entire result of the
* append node may not be scanned, but this way all of the
* structures get allocated in the executor's top level memory
* block instead of that of the call to ExecAppend.)
* (這可能是一種浪費，因為可能不會掃描append節點的整個結果，
* 但是透過這種方式，所有結構都將在執行程式的最頂級記憶體塊中分配，
* 而不是在ExecAppend呼叫的記憶體塊中分配。)
* ----------------------------------------------------------------
*/
AppendState *
ExecInitAppend(Append *node, EState *estate, int eflags)
{
AppendState *appendstate = makeNode(AppendState);
PlanState **appendplanstates;
Bitmapset *validsubplans;
int nplans;
int firstvalid;
int i,
j;
ListCell *lc;
/* check for unsupported flags */
//檢查不支援的標記
Assert(!(eflags & EXEC_FLAG_MARK));
/*
* create new AppendState for our append node
* 為append節點建立AppendState
*/
appendstate->ps.plan = (Plan *) node;
appendstate->ps.state = estate;
appendstate->ps.ExecProcNode = ExecAppend;
/* Let choose_next_subplan_* function handle setting the first subplan */
//讓choose_next_subplan_* 函式處理第一個子計劃的設定
appendstate->as_whichplan = INVALID_SUBPLAN_INDEX;
/* If run-time partition pruning is enabled, then set that up now */
//如允許執行時分割槽裁剪,則配置分割槽裁剪
if (node->part_prune_info != NULL)
{
PartitionPruneState *prunestate;
/* We may need an expression context to evaluate partition exprs */
//需要獨立的表示式上下文來解析分割槽表示式
ExecAssignExprContext(estate, &appendstate->ps);
/* Create the working data structure for pruning. */
//為裁剪建立工作資料結構資訊
prunestate = ExecCreatePartitionPruneState(&appendstate->ps,
node->part_prune_info);
appendstate->as_prune_state = prunestate;
/* Perform an initial partition prune, if required. */
//如需要,執行初始的分割槽裁剪
if (prunestate->do_initial_prune)
{
/* Determine which subplans survive initial pruning */
//確定那個子計劃在初始裁剪中仍"存活"
validsubplans = ExecFindInitialMatchingSubPlans(prunestate,
list_length(node->appendplans));
/*
* The case where no subplans survive pruning must be handled
* specially. The problem here is that code in explain.c requires
* an Append to have at least one subplan in order for it to
* properly determine the Vars in that subplan's targetlist. We
* sidestep this issue by just initializing the first subplan and
* setting as_whichplan to NO_MATCHING_SUBPLANS to indicate that
* we don't really need to scan any subnodes.
* 沒有子計劃"存活"的情況需要特別處理.
* 這裡的問題是explain.c的執行邏輯至少需要一個Append節點和一個子計劃,
* 用於確定子計劃投影列連結串列中的Vars.
* 透過初始化第一個子計劃以及設定as_whichplan為NO_MATCHING_SUBPLANS,
* 用以指示不需要掃描任何的子節點.
*/
if (bms_is_empty(validsubplans))
{
appendstate->as_whichplan = NO_MATCHING_SUBPLANS;
/* Mark the first as valid so that it's initialized below */
//標記第一個子計劃為有效的,以便在接下來可以初始化
validsubplans = bms_make_singleton(0);
}
//子計劃數目
nplans = bms_num_members(validsubplans);
}
else
{
/* We'll need to initialize all subplans */
//需要初始化所有的的子計劃
nplans = list_length(node->appendplans);
Assert(nplans > 0);
validsubplans = bms_add_range(NULL, 0, nplans - 1);
}
/*
* If no runtime pruning is required, we can fill as_valid_subplans
* immediately, preventing later calls to ExecFindMatchingSubPlans.
* 如不需要執行期裁剪,那麼我們可以馬上設定as_valid_subplans,
* 以防止後續邏輯呼叫ExecFindMatchingSubPlans
*/
if (!prunestate->do_exec_prune)
{
Assert(nplans > 0);
appendstate->as_valid_subplans = bms_add_range(NULL, 0, nplans - 1);
}
}
else
{
//不需要執行期裁剪
nplans = list_length(node->appendplans);
/*
* When run-time partition pruning is not enabled we can just mark all
* subplans as valid; they must also all be initialized.
* 執行期裁剪不允許,設定所有的子計劃為有效,同時這些子計劃必須全部初始化
*/
Assert(nplans > 0);
appendstate->as_valid_subplans = validsubplans =
bms_add_range(NULL, 0, nplans - 1);
appendstate->as_prune_state = NULL;
}
/*
* Initialize result tuple type and slot.
* 初始化結果元組型別和slot
*/
ExecInitResultTupleSlotTL(&appendstate->ps, &TTSOpsVirtual);
/* node returns slots from each of its subnodes, therefore not fixed */
//從每一個子節點中返回slot,但並不固定
appendstate->ps.resultopsset = true;
appendstate->ps.resultopsfixed = false;
appendplanstates = (PlanState **) palloc(nplans *
sizeof(PlanState *));
/*
* call ExecInitNode on each of the valid plans to be executed and save
* the results into the appendplanstates array.
* 在每一個有效的計劃上執行ExecInitNode,
* 同時儲存結果到appendplanstates陣列中.
*
* While at it, find out the first valid partial plan.
* 在此期間,找到第一個有效的部分計劃(並行使用)
*/
j = i = 0;
firstvalid = nplans;
foreach(lc, node->appendplans)
{
if (bms_is_member(i, validsubplans))
{
Plan *initNode = (Plan *) lfirst(lc);
/*
* Record the lowest appendplans index which is a valid partial
* plan.
* 記錄有效部分計劃的最小appendplans索引號
*/
if (i >= node->first_partial_plan && j < firstvalid)
firstvalid = j;
appendplanstates[j++] = ExecInitNode(initNode, estate, eflags);//初始化節點
}
i++;
}
//配置Append State
appendstate->as_first_partial_plan = firstvalid;
appendstate->appendplans = appendplanstates;
appendstate->as_nplans = nplans;
/*
* Miscellaneous initialization
* 初始化
*/
appendstate->ps.ps_ProjInfo = NULL;
/* For parallel query, this will be overridden later. */
//對於並行查詢,該值後面會被覆蓋
appendstate->choose_next_subplan = choose_next_subplan_locally;
return appendstate;
}
ExecAppend
ExecAppend在多個子計劃中進行迭代處理
/* ----------------------------------------------------------------
* ExecAppend
*
* Handles iteration over multiple subplans.
* 在多個子計劃中處理迭代
* ----------------------------------------------------------------
*/
static TupleTableSlot *
ExecAppend(PlanState *pstate)
{
AppendState *node = castNode(AppendState, pstate);
if (node->as_whichplan < 0)
{
/*
* If no subplan has been chosen, we must choose one before
* proceeding.
* 如果沒有子計劃選中,必須在開始前選擇一個
*/
if (node->as_whichplan == INVALID_SUBPLAN_INDEX &&
!node->choose_next_subplan(node))
return ExecClearTuple(node->ps.ps_ResultTupleSlot);//出錯,則清除tuple,返回
/* Nothing to do if there are no matching subplans */
//如無匹配的子計劃,返回
else if (node->as_whichplan == NO_MATCHING_SUBPLANS)
return ExecClearTuple(node->ps.ps_ResultTupleSlot);
}
for (;;)//迴圈獲取tuple slot
{
PlanState *subnode;
TupleTableSlot *result;
CHECK_FOR_INTERRUPTS();
/*
* figure out which subplan we are currently processing
* 選擇哪個子計劃需要現在執行
*/
Assert(node->as_whichplan >= 0 && node->as_whichplan < node->as_nplans);
subnode = node->appendplans[node->as_whichplan];
/*
* get a tuple from the subplan
* 從子計劃中獲取tuple
*/
result = ExecProcNode(subnode);
if (!TupIsNull(result))
{
/*
* If the subplan gave us something then return it as-is. We do
* NOT make use of the result slot that was set up in
* ExecInitAppend; there's no need for it.
* 如果子計劃有返回,則直接返回.
* 在這裡,不需要在ExecInitAppend中構造結果slot,因為不需要這樣做.
*/
return result;
}
/* choose new subplan; if none, we're done */
//選擇新的子計劃,如無,則完成呼叫
if (!node->choose_next_subplan(node))
return ExecClearTuple(node->ps.ps_ResultTupleSlot);
}
}
三、跟蹤分析
測試指令碼如下
testdb=# explain verbose select * from t_hash_partition where c1 = 1 OR c1 = 2;
QUERY PLAN
-------------------------------------------------------------------------------------
Append (cost=0.00..30.53 rows=6 width=200)
-> Seq Scan on public.t_hash_partition_1 (cost=0.00..15.25 rows=3 width=200)
Output: t_hash_partition_1.c1, t_hash_partition_1.c2, t_hash_partition_1.c3
Filter: ((t_hash_partition_1.c1 = 1) OR (t_hash_partition_1.c1 = 2))
-> Seq Scan on public.t_hash_partition_3 (cost=0.00..15.25 rows=3 width=200)
Output: t_hash_partition_3.c1, t_hash_partition_3.c2, t_hash_partition_3.c3
Filter: ((t_hash_partition_3.c1 = 1) OR (t_hash_partition_3.c1 = 2))
(7 rows)
ExecInitAppend
啟動gdb,設定斷點
(gdb) b ExecInitAppend
Breakpoint 1 at 0x6efa8a: file nodeAppend.c, line 103.
(gdb) c
Continuing.
Breakpoint 1, ExecInitAppend (node=0x27af638, estate=0x27be058, eflags=16) at nodeAppend.c:103
103 AppendState *appendstate = makeNode(AppendState);
初始化AppendState
(gdb) n
113 Assert(!(eflags & EXEC_FLAG_MARK));
(gdb)
119 ExecLockNonLeafAppendTables(node->partitioned_rels, estate);
(gdb)
124 appendstate->ps.plan = (Plan *) node;
(gdb)
125 appendstate->ps.state = estate;
(gdb)
126 appendstate->ps.ExecProcNode = ExecAppend;
(gdb)
129 appendstate->as_whichplan = INVALID_SUBPLAN_INDEX;
(gdb)
不需要執行期分割槽裁剪
(gdb)
132 if (node->part_prune_info != NULL)
(gdb) p node->part_prune_info
$1 = (struct PartitionPruneInfo *) 0x0
(gdb)
需要初始化所有的的子計劃
(gdb) n
190 nplans = list_length(node->appendplans);
(gdb)
196 Assert(nplans > 0);
(gdb)
198 bms_add_range(NULL, 0, nplans - 1);
(gdb)
197 appendstate->as_valid_subplans = validsubplans =
(gdb) n
199 appendstate->as_prune_state = NULL;
(gdb) p *validsubplans
$4 = {nwords = 1, words = 0x27be38c}
(gdb) p *validsubplans->words
$5 = 3 --> 即No.0 + No.1
(gdb)
初始化結果元組型別和slot
(gdb) n
205 ExecInitResultTupleSlotTL(estate, &appendstate->ps);
(gdb)
207 appendplanstates = (PlanState **) palloc(nplans *
(gdb)
在每一個有效的計劃上執行ExecInitNode,同時保持結果到appendplanstates陣列中.
(gdb)
216 j = i = 0;
(gdb) n
217 firstvalid = nplans;
(gdb)
218 foreach(lc, node->appendplans)
(gdb) p nplans
$6 = 2
(gdb) p node->appendplans
$7 = (List *) 0x27b30f0
(gdb) p *node->appendplans
$8 = {type = T_List, length = 2, head = 0x27b30c8, tail = 0x27b33d8}
(gdb)
遍歷appendplans,初始化appendplans中的節點(SeqScan)
(gdb) n
220 if (bms_is_member(i, validsubplans))
(gdb) n
222 Plan *initNode = (Plan *) lfirst(lc);
(gdb)
228 if (i >= node->first_partial_plan && j < firstvalid)
(gdb)
231 appendplanstates[j++] = ExecInitNode(initNode, estate, eflags);
(gdb) p j
$9 = 0
(gdb) p i
$10 = 0
(gdb) n
233 i++;
(gdb)
218 foreach(lc, node->appendplans)
(gdb)
220 if (bms_is_member(i, validsubplans))
(gdb)
222 Plan *initNode = (Plan *) lfirst(lc);
(gdb)
228 if (i >= node->first_partial_plan && j < firstvalid)
(gdb) p *initNode
$11 = {type = T_SeqScan, startup_cost = 0, total_cost = 15.25, plan_rows = 3, plan_width = 200, parallel_aware = false,
parallel_safe = true, plan_node_id = 2, targetlist = 0x27b31a8, qual = 0x27b3308, lefttree = 0x0, righttree = 0x0,
initPlan = 0x0, extParam = 0x0, allParam = 0x0}
(gdb) n
231 appendplanstates[j++] = ExecInitNode(initNode, estate, eflags);
(gdb)
233 i++;
(gdb)
218 foreach(lc, node->appendplans)
(gdb)
236 appendstate->as_first_partial_plan = firstvalid;
(gdb)
完成初始化,其中choose_next_subplan函式為choose_next_subplan_locally函式
(gdb) p firstvalid
$12 = 2
(gdb) n
237 appendstate->appendplans = appendplanstates;
(gdb)
238 appendstate->as_nplans = nplans;
(gdb)
244 appendstate->ps.ps_ProjInfo = NULL;
(gdb)
247 appendstate->choose_next_subplan = choose_next_subplan_locally;
(gdb)
249 return appendstate;
(gdb) p choose_next_subplan_locally
$13 = {_Bool (AppendState *)} 0x6f02d8 <choose_next_subplan_locally>
(gdb) p *appendstate
$15 = {ps = {type = T_AppendState, plan = 0x27af638, state = 0x27be058, ExecProcNode = 0x6efe19 <ExecAppend>,
ExecProcNodeReal = 0x0, instrument = 0x0, worker_instrument = 0x0, worker_jit_instrument = 0x0, qual = 0x0,
lefttree = 0x0, righttree = 0x0, initPlan = 0x0, subPlan = 0x0, chgParam = 0x0, ps_ResultTupleSlot = 0x27be5c0,
ps_ExprContext = 0x0, ps_ProjInfo = 0x0, scandesc = 0x0}, appendplans = 0x27be6b8, as_nplans = 2, as_whichplan = -1,
as_first_partial_plan = 2, as_pstate = 0x0, pstate_len = 0, as_prune_state = 0x0, as_valid_subplans = 0x27be388,
choose_next_subplan = 0x6f02d8 <choose_next_subplan_locally>}
ExecAppend
設定斷點,進入ExecAppend
(gdb) del
Delete all breakpoints? (y or n) y
(gdb) b ExecAppend
Breakpoint 2 at 0x6efe2a: file nodeAppend.c, line 261.
(gdb) c
Continuing.
Breakpoint 2, ExecAppend (pstate=0x27be270) at nodeAppend.c:261
261 AppendState *node = castNode(AppendState, pstate);
輸入引數為先前已完成初始化的AppendState
(gdb) p *(AppendState *)pstate
$19 = {ps = {type = T_AppendState, plan = 0x27af638, state = 0x27be058, ExecProcNode = 0x6efe19 <ExecAppend>,
ExecProcNodeReal = 0x6efe19 <ExecAppend>, instrument = 0x0, worker_instrument = 0x0, worker_jit_instrument = 0x0,
qual = 0x0, lefttree = 0x0, righttree = 0x0, initPlan = 0x0, subPlan = 0x0, chgParam = 0x0,
ps_ResultTupleSlot = 0x27be5c0, ps_ExprContext = 0x0, ps_ProjInfo = 0x0, scandesc = 0x0}, appendplans = 0x27be6b8,
as_nplans = 2, as_whichplan = -1, as_first_partial_plan = 2, as_pstate = 0x0, pstate_len = 0, as_prune_state = 0x0,
as_valid_subplans = 0x27be388, choose_next_subplan = 0x6f02d8 <choose_next_subplan_locally>}
如果沒有子計劃選中,必須在開始前選擇一個(選擇子計劃:No.0)
(gdb) n
263 if (node->as_whichplan < 0)
(gdb)
269 if (node->as_whichplan == INVALID_SUBPLAN_INDEX &&
(gdb)
270 !node->choose_next_subplan(node))
(gdb)
269 if (node->as_whichplan == INVALID_SUBPLAN_INDEX &&
(gdb)
274 else if (node->as_whichplan == NO_MATCHING_SUBPLANS)
(gdb) p node->as_whichplan
$20 = 0
(gdb) n
283 CHECK_FOR_INTERRUPTS();
獲取子計劃並執行
(gdb)
288 Assert(node->as_whichplan >= 0 && node->as_whichplan < node->as_nplans);
(gdb)
289 subnode = node->appendplans[node->as_whichplan];
(gdb)
294 result = ExecProcNode(subnode);
(gdb) p *subnode
$21 = {type = T_SeqScanState, plan = 0x27b2728, state = 0x27be058, ExecProcNode = 0x6e4bde <ExecProcNodeFirst>,
ExecProcNodeReal = 0x71578d <ExecSeqScan>, instrument = 0x0, worker_instrument = 0x0, worker_jit_instrument = 0x0,
qual = 0x27bec88, lefttree = 0x0, righttree = 0x0, initPlan = 0x0, subPlan = 0x0, chgParam = 0x0,
ps_ResultTupleSlot = 0x27bebc8, ps_ExprContext = 0x27be7f8, ps_ProjInfo = 0x0, scandesc = 0x7f7d049417e0}
返回slot
(gdb) n
296 if (!TupIsNull(result))
(gdb) p *result
$22 = {type = T_TupleTableSlot, tts_isempty = false, tts_shouldFree = false, tts_shouldFreeMin = false, tts_slow = false,
tts_tuple = 0x27c5500, tts_tupleDescriptor = 0x7f7d049417e0, tts_mcxt = 0x27bdf40, tts_buffer = 120, tts_nvalid = 1,
tts_values = 0x27be950, tts_isnull = 0x27be968, tts_mintuple = 0x0, tts_minhdr = {t_len = 0, t_self = {ip_blkid = {
bi_hi = 0, bi_lo = 0}, ip_posid = 0}, t_tableOid = 0, t_data = 0x0}, tts_off = 4, tts_fixedTupleDescriptor = true}
如第一個子計劃執行完畢,則選擇下一個子計劃(呼叫choose_next_subplan函式切換到下一個子計劃)
(gdb) c
Continuing.
Breakpoint 2, ExecAppend (pstate=0x27be270) at nodeAppend.c:261
261 AppendState *node = castNode(AppendState, pstate);
(gdb) n
263 if (node->as_whichplan < 0)
(gdb)
283 CHECK_FOR_INTERRUPTS();
(gdb)
288 Assert(node->as_whichplan >= 0 && node->as_whichplan < node->as_nplans);
(gdb)
289 subnode = node->appendplans[node->as_whichplan];
(gdb)
294 result = ExecProcNode(subnode);
(gdb)
296 if (!TupIsNull(result))
(gdb)
307 if (!node->choose_next_subplan(node))
(gdb)
309 }
DONE!
四、參考資料
Parallel Append implementation
Partition Elimination in PostgreSQL 11
來自 “ ITPUB部落格 ” ，連結：http://blog.itpub.net/6906/viewspace-2374787/，如需轉載，請註明出處，否則將追究法律責任。
相關文章
- PostgreSQL 原始碼解讀（98）- 分割槽表#4（資料查詢路由#1-“擴充套件”分割槽表）SQL原始碼路由套件
- PostgreSQL 原始碼解讀（99）- 分割槽表#5（資料查詢路由#2-RelOptInfo數...SQL原始碼路由
- PostgreSQL 原始碼解讀（100）- 分割槽表#6（資料查詢路由#3-prune part...SQL原始碼路由
- PostgreSQL 原始碼解讀（92）- 分割槽表#1（資料插入路由#1）SQL原始碼路由
- PostgreSQL 原始碼解讀（94）- 分割槽表#2（資料插入路由#2）SQL原始碼路由
- PostgreSQL 原始碼解讀（101）- 分割槽表#7（資料查詢路由#4-prune part...SQL原始碼路由
- PostgreSQL 原始碼解讀（96）- 分割槽表#3（資料插入路由#3-獲取分割槽鍵值）SQL原始碼路由
- PostgreSQL 原始碼解讀（102）- 分割槽表#8（資料查詢路由#5-構建APPEND訪問路徑）SQL原始碼路由APP
- PostgreSQL 原始碼解讀（24）- 查詢語句#9（查詢重寫）SQL原始碼
- PostgreSQL 原始碼解讀（187）- 查詢#103(聚合函式#8 - Struct Review)SQL原始碼函式StructView
- postgresql分割槽表實現方式SQL
- PostgreSQL 原始碼解讀（234）- 查詢#127(NOT IN實現#5)SQL原始碼
- PostgreSQL 原始碼解讀（232）- 查詢#125(NOT IN實現#3)SQL原始碼
- PostgreSQL 原始碼解讀（231）- 查詢#124(NOT IN實現#2)SQL原始碼
- PostgreSQL 原始碼解讀（230）- 查詢#123(NOT IN實現)SQL原始碼
- oracle實用sql(14)--查詢分割槽表的分割槽列和子分割槽列OracleSQL
- PostgreSQL 原始碼解讀（10）- 插入資料#9（ProcessQuery）SQL原始碼
- PostgreSQL 原始碼解讀（9）- 插入資料#8（ExecutorRun和standard...SQL原始碼
- PostgreSQL使用表繼承實現分割槽表SQL繼承
- Oracle查詢Interval partition分割槽表內資料Oracle
- PostgreSQL 原始碼解讀（233）- 查詢#126(NOT IN實現#4)SQL原始碼
- PostgreSQL 原始碼解讀（194）- 查詢#110(排序#3 - 實現)SQL原始碼排序
- PostgreSQL 原始碼解讀（195）- 查詢#111(排序#4 - 實現)SQL原始碼排序
- PostgreSQL 原始碼解讀（160）- 查詢#80(如何實現表示式解析)SQL原始碼
- 分割槽表分割槽索引查詢效率探究索引
- 如何查詢分割槽表的分割槽及子分割槽
- postgresql分割槽表修改資料表欄位SQL
- Oracle查詢資料庫中所有表和分割槽表的記錄數Oracle資料庫
- postgresql分割槽表SQL
- PostgreSQL 原始碼解讀（48）- 查詢語句#33（query_planner函式#9）SQL原始碼函式
- PostgreSQL 原始碼解讀（20）- 查詢語句#5（查詢樹Query詳解）SQL原始碼
- PostgreSQL 原始碼解讀（203）- 查詢#116(型別轉換實現)SQL原始碼型別
- PostgreSQL 原始碼解讀（205）- 查詢#118(資料結構RangeTblEntry)SQL原始碼資料結構
- PostgreSQL和oracle表分割槽對比SQLOracle
- PostgreSQL 原始碼解讀（19）- 查詢語句#4（ParseTree詳解）SQL原始碼
- PostgreSQL 原始碼解讀（25）- 查詢語句#10（查詢優化概覽）SQL原始碼優化
- PostgreSQL 原始碼解讀（17）- 查詢語句#2（查詢優化基礎）SQL原始碼優化
- PostgreSQL 原始碼解讀（164）- 查詢#84(表示式求值)SQL原始碼