enh(interp): fill(prev/next/near/linear) is configured by ignoreNull param #34006

Tony2h · 2025-12-22T07:15:06Z

Description

Issue(s)

resolve: https://project.feishu.cn/taosdata_td/feature/detail/6506970855

Checklist

Please check the items in the checklist if applicable.

Is the user manual updated?
Are the test cases passed and automated?
Is there no significant decrease in test coverage?

Tony2h · 2025-12-22T07:15:30Z

/gemini review

gemini-code-assist · 2025-12-22T07:15:39Z

Summary of Changes

Hello @Tony2h, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request significantly improves the flexibility and efficiency of interpolation functions by introducing explicit null handling and a robust operator notification system. The changes allow interpolation queries to better manage null values and enable operators to communicate their processing progress, leading to optimized data fetching and reduced overhead in complex query scenarios. The integration spans message structures, executor logic, storage API, and various operator implementations, ensuring a cohesive enhancement to the system's data processing capabilities.

Highlights

Enhanced Interpolation with Null Handling: The interpolation functions (fill(prev/next/near/linear)) now support an ignoreNull parameter, allowing more granular control over how null values are treated during data filling. This refines the logic for isInvalidTimestamp and checkNullRow within the timeslice operator.
New Task Notification Mechanism: A new TASK_NOTIFY_STEP_DONE type has been introduced for task notifications, along with a pOpParam field in STaskNotifyReq to carry operator-specific parameters. This enables operators to communicate completion of specific processing steps.
Operator Notification System Integration: A comprehensive notification system has been integrated across various components. This includes new serialization/deserialization functions for operator parameters, a notifyTableScanTask in the executor, and a tsdReaderStepDone function in the storage API to mark reader steps as complete. Scan and Exchange operators now have registered notification handlers.
Optimized Reader Capacity and Step Management: Inner readers for timeseries data now initialize with a larger capacity (1024 instead of 1) to reduce reallocations. The reader state machine has been enhanced with EXTERNAL_ROWS_INIT and EXTERNAL_ROWS_DONE states, and refined logic for transitioning between PREV, MAIN, and NEXT scan steps using a currentStepDone flag.
New Test Cases for Interpolation: New test cases (interp_fill_ignore_null.in and .csv) have been added, along with corresponding Python test script modifications, to validate the enhanced interpolation functionality, especially concerning null value handling.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	`/gemini review`	Performs a code review for the current pull request in its current state.
Pull Request Summary	`/gemini summary`	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	`/gemini help`	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

gemini-code-assist

Code Review

This pull request introduces a mechanism for the interp operator to notify downstream scan operators to optimize data fetching for fill operations. This is achieved by adding a new task notification type and associated serialization/deserialization logic. The changes span across messaging, executor, and storage layers. While the overall approach is sound, there are several critical memory leaks in the new message deserialization and handling logic that must be addressed. I've also found a potential logic bug in a condition check and some leftover debug code. Please review my comments for details.

source/common/src/msg/tmsg.c

+static int32_t tDeserializeSOperatorNotifyParam(SDecoder *pDecoder, SOperatorParam *pOpParam) {
+  TAOS_CHECK_RETURN(tDecodeI32(pDecoder, &pOpParam->opType));
+  TAOS_CHECK_RETURN(tDecodeI32(pDecoder, &pOpParam->downstreamIdx));
+  switch (pOpParam->opType) {
+    case QUERY_NODE_PHYSICAL_PLAN_TABLE_SCAN: {
+      pOpParam->value = taosMemoryMalloc(sizeof(int64_t));
+      if (NULL == pOpParam->value) {
+        TAOS_CHECK_RETURN(terrno);
+      }
+      TAOS_CHECK_RETURN(tDecodeI64(pDecoder, (int64_t*)pOpParam->value));
+      break;
+    }
+    default:
+      return TSDB_CODE_INVALID_PARA;
+  }
+
+  int32_t childrenNum = 0;
+  TAOS_CHECK_RETURN(tDecodeI32(pDecoder, &childrenNum));
+
+  if (childrenNum > 0) {
+    pOpParam->pChildren = taosArrayInit(childrenNum, POINTER_BYTES);
+    if (NULL == pOpParam->pChildren) {
+      TAOS_CHECK_RETURN(terrno);
+    }
+    for (int32_t i = 0; i < childrenNum; ++i) {
+      SOperatorParam *pChild = taosMemoryCalloc(1, sizeof(SOperatorParam));
+      if (NULL == pChild) {
+        TAOS_CHECK_RETURN(terrno);
+      }
+      TAOS_CHECK_RETURN(tDeserializeSOperatorNotifyParam(pDecoder, pChild));
+      if (taosArrayPush(pOpParam->pChildren, &pChild) == NULL) {
+        TAOS_CHECK_RETURN(terrno);
+      }
+    }
+  } else {
+    pOpParam->pChildren = NULL;
+  }
+
+  TAOS_CHECK_RETURN(tDecodeBool(pDecoder, &pOpParam->reUse));
+
+  return TSDB_CODE_SUCCESS;
+}


source/common/src/msg/tmsg.c

+    pReq->pOpParam = taosMemoryCalloc(1, sizeof(SOperatorParam));
+    if (NULL == pReq->pOpParam) {
+      TAOS_CHECK_EXIT(terrno);
+    }
+    TAOS_CHECK_EXIT(tDeserializeSOperatorNotifyParam(&decoder, pReq->pOpParam));


source/libs/qworker/src/qworker.c

+    case TASK_NOTIFY_STEP_DONE: {
+      if (ctx->taskHandle != NULL) {
+        QW_ERR_JRET(
+          notifyTableScanTask(ctx->taskHandle,
+                              ((STaskNotifyReq*)qwMsg->msg)->pOpParam));
+      }
+      break;
+    }


source/libs/executor/src/timesliceoperator.c


-    if (checkNullRow(&pOperator->exprSupp, pBlock, i, ignoreNull)) {
-      continue;
+    if (!notified && (ts < pSliceInfo->win.ekey || ts > pSliceInfo->win.skey)) {


source/dnode/vnode/src/tsdb/tsdbRead2.c

+/*
+  @brief Mark the current step done, so next step will be triggered.
+  @param pReader the reader to mark the step done
+  @param notifyTs the timestamp to notify, used to determine whether  


source/libs/planner/src/planSpliter.c

+      // TODO(tony): for debug!!!!!!!!!!!
+      // QUERY_NODE_LOGIC_PLAN_INTERP_FUNC != nodeType(pNode->pParent) &&


Copilot

Pull request overview

This pull request enhances the interpolation (interp) functionality for fill operations (prev/next/near/linear) to be configured by an ignoreNull parameter. The implementation introduces a notification mechanism that allows the time slice operator to notify downstream table scan operators when a step is complete, optimizing the prev/next scan behavior to avoid unnecessary data fetching.

Key changes include:

Implementation of a step-done notification system for coordinating between interpolation and scan operators
Refactoring of timestamp validation logic to properly handle null values during interpolation
Extension of reader state management to track completion of prev/next/main scan steps

Reviewed changes

Copilot reviewed 22 out of 22 changed files in this pull request and generated 12 comments.

Show a summary per file

File	Description
test/cases/09-DataQuerying/12-Interp/test_query_interp_fill.py	Adds test setup with new tables and test function for interpolation with null value handling
test/cases/09-DataQuerying/12-Interp/in/interp_fill_ignore_null.in	Test queries exercising fill(prev/next/linear) with various range and partition scenarios
test/cases/09-DataQuerying/12-Interp/ans/interp_fill_ignore_null.csv	Expected results showing proper null handling in interpolation
source/libs/executor/src/timesliceoperator.c	Core changes: refactored timestamp validation, added downstream notification logic, simplified loop structure
source/libs/executor/src/scanoperator.c	Implements notification handler for table scan operator to mark reader step done
source/libs/executor/src/exchangeoperator.c	Adds notification forwarding to remote sources via RPC messages
source/libs/executor/src/executor.c	Entry point for notifying table scan tasks from worker nodes
source/libs/executor/src/operator.c	Utility function to set notification callback on operators
source/libs/executor/src/executorInt.c	Parameter building and cleanup for notification messages
source/dnode/vnode/src/tsdb/tsdbRead2.c	Enhanced reader state machine with step tracking and notification handling
source/dnode/vnode/src/tsdb/tsdbReadUtil.h	Added EXTERNAL_ROWS_INIT and EXTERNAL_ROWS_DONE states, currentStepDone flag
source/dnode/vnode/src/vnd/vnodeApi.c	Registers reader step done callback in API structure
source/libs/qworker/src/qworker.c	Handles TASK_NOTIFY_STEP_DONE message type in worker
source/libs/qworker/src/qwMsg.c	Passes notify message payload to worker processing
source/common/src/msg/tmsg.c	Separates serialization for GET vs NOTIFY operator parameters
source/libs/planner/src/planSpliter.c	Temporarily comments out interp function split check (marked as debug code)
include/libs/executor/*.h	Function declarations for notification infrastructure
include/common/tmsg.h	Adds TASK_NOTIFY_STEP_DONE type and pOpParam field
include/libs/executor/storageapi.h	Adds tsdReaderStepDone callback to TsdReader API

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

source/libs/executor/src/exchangeoperator.c

+      continue;
+    }
+
+    // TODO(Tianyi Zhang): optimize localExec


source/libs/executor/src/timesliceoperator.c

+    if (!notified && (ts < pSliceInfo->win.ekey || ts > pSliceInfo->win.skey)) {
+      /*
+        When downstream contains table scan operator, it may do prev/next
+        scan to fill(prev/next/linear/near) the first/last row. To reduce 
+        the overhead, notify it only once when the first valid row comes.
+      */
+      code = timeSliceOptrNotifyDownstream(pOperator->pDownstream[0], ts);
+      QUERY_CHECK_CODE(code, lino, _end);
+      notified = true;


source/libs/executor/src/exchangeoperator.c

+    code = buildOperatorStepDoneNotifyParam(&pNotifyParam, 
+                                            QUERY_NODE_PHYSICAL_PLAN_TABLE_SCAN,
+                                            *(int64_t*)param->value);
+    if (code != TSDB_CODE_SUCCESS || pNotifyParam == NULL) {
+      qError("%s exchange optr failed to build notify param for source %d, "
+             "since:%s", GET_TASKID(pTaskInfo), i, tstrerror(code));
+      freeOperatorParam(pNotifyParam, OP_NOTIFY_PARAM);
+      continue;
+    }


source/dnode/vnode/src/tsdb/tsdbRead2.c

+  @brief Mark the current step done, so next step will be triggered.
+  @param pReader the reader to mark the step done
+  @param notifyTs the timestamp to notify, used to determine whether  
+*/


source/libs/executor/src/timesliceoperator.c

+static bool isInvalidTimestamp(STimeSliceOperatorInfo* pSliceInfo,
+                                   int64_t currentTs, SColumnInfoData* pPkCol,
+                                   int32_t curIndex) {
  if (currentTs > pSliceInfo->win.ekey) {
    return false;
  }
+  if (pSliceInfo->prevTsSet && currentTs <= pSliceInfo->prevKey.ts) {
+    /*
+      Input data of time slice operator must be ordered by
+      timestamp ascendingly, except the prev scan.
+      So prevTs should never be updated to equal or smaller timestamp.
+    */
+    return true;
+  }


source/libs/executor/src/timesliceoperator.c


-    if (checkNullRow(&pOperator->exprSupp, pBlock, i, ignoreNull)) {
-      continue;
+    if (!notified && (ts < pSliceInfo->win.ekey || ts > pSliceInfo->win.skey)) {


source/libs/executor/src/timesliceoperator.c

+  @param pDownOptr the downstream operator info
+  @param pInput the input parameter for the notify function
+  @return the code of the operation
+*/
+static int32_t timeSliceOptrNotifyDownstream(SOperatorInfo* pDownOptr,
+                                             int64_t notifyTs) {
+  int32_t code = TSDB_CODE_SUCCESS;
+  int32_t lino = 0;
+  SOperatorParam* pNotifyParam = NULL;
+  if (pDownOptr != NULL && pDownOptr->fpSet.notifyFn != NULL) {
+    code = buildOperatorStepDoneNotifyParam(&pNotifyParam,
+                                            pDownOptr->operatorType, notifyTs);
+    QUERY_CHECK_CODE(code, lino, _end);
+
+    code = pDownOptr->fpSet.notifyFn(pDownOptr, pNotifyParam);


source/libs/executor/src/timesliceoperator.c

  SSDataBlock* pResBlock = pSliceInfo->pRes;
  SInterval*   pInterval = &pSliceInfo->interval;
+  bool         notified = false;
+  SOperatorParam* pNotifyParam = NULL;


source/libs/executor/src/exchangeoperator.c

+    void* msg = taosMemoryCalloc(1, msgSize);
+    if (msg == NULL) {
+      qError("%s failed to alloc memory for notify msg, size:%d", 
+             GET_TASKID(pTaskInfo), msgSize);
+      freeOperatorParam(pNotifyParam, OP_NOTIFY_PARAM);
+      continue;
+    }
+
+    msgSize = tSerializeSTaskNotifyReq(msg, msgSize, &req);
+    if (msgSize < 0) {
+      qError("%s failed to serialize notify req for source %d, size:%d", 
+             GET_TASKID(pTaskInfo), i, msgSize);
+      taosMemoryFree(msg);
+      freeOperatorParam(pNotifyParam, OP_NOTIFY_PARAM);
+      continue;
+    }
+
+    // send notify msg to source
+    SMsgSendInfo* pMsgSendInfo = taosMemoryCalloc(1, sizeof(SMsgSendInfo));
+    if (pMsgSendInfo == NULL) {
+      qError("%s failed to alloc memory for msg send info",
+             GET_TASKID(pTaskInfo));
+      taosMemoryFree(msg);
+      freeOperatorParam(pNotifyParam, OP_NOTIFY_PARAM);
+      continue;
+    }
+
+    pMsgSendInfo->msgInfo.pData = msg;
+    pMsgSendInfo->msgInfo.len = msgSize;
+    pMsgSendInfo->msgType = TDMT_SCH_TASK_NOTIFY;
+    pMsgSendInfo->requestId = pTaskInfo->id.queryId;
+    pMsgSendInfo->fp = NULL;
+    pMsgSendInfo->param = NULL;
+    pMsgSendInfo->paramFreeFp = NULL;
+
+    int64_t transporterId = 0;
+    code = asyncSendMsgToServer(pExchangeInfo->pTransporter,
+                                &pSource->addr.epSet, &transporterId,
+                                pMsgSendInfo);
+    if (code != TSDB_CODE_SUCCESS) {
+      qError("%s failed to send notify msg to source %d (vgId:%d), code:%s", 
+             GET_TASKID(pTaskInfo), i, pSource->addr.nodeId, tstrerror(code));
+      taosMemoryFree(msg);
+      taosMemoryFree(pMsgSendInfo);
+      freeOperatorParam(pNotifyParam, OP_NOTIFY_PARAM);
+      continue;
+    }


source/dnode/vnode/src/tsdb/tsdbRead2.c


-  if (pReader->innerReader[0] != NULL && pReader->step == 0) {
+  if (pReader->innerReader[0] != NULL &&
+      pReader->step == EXTERNAL_ROWS_PREV &&


gemini-code-assist

Code Review

This pull request introduces an enhancement for interp queries with fill by adding a notification mechanism to optimize scanning. This allows the timeslice operator to inform downstream scan operators to proceed to the next scanning phase, which is a good optimization. The changes are extensive, touching message serialization, various query operators, and the storage layer. While the overall approach is sound and new tests are included, I've identified several critical memory leaks in the new deserialization logic due to missing error handling. Additionally, there are some minor issues, including leftover debug code, an incomplete comment, and potential for unhelpful error messages due to incorrect line number reporting.

source/common/src/msg/tmsg.c

+      pOpParam->value = taosMemoryMalloc(sizeof(int64_t));
+      if (NULL == pOpParam->value) {
+        TAOS_CHECK_RETURN(terrno);
+      }
+      TAOS_CHECK_RETURN(tDecodeI64(pDecoder, (int64_t*)pOpParam->value));
+      break;


source/common/src/msg/tmsg.c

+    for (int32_t i = 0; i < childrenNum; ++i) {
+      SOperatorParam *pChild = taosMemoryCalloc(1, sizeof(SOperatorParam));
+      if (NULL == pChild) {
+        TAOS_CHECK_RETURN(terrno);
+      }
+      TAOS_CHECK_RETURN(tDeserializeSOperatorNotifyParam(pDecoder, pChild));
+      if (taosArrayPush(pOpParam->pChildren, &pChild) == NULL) {
+        TAOS_CHECK_RETURN(terrno);
+      }
+    }


source/common/src/msg/tmsg.c

  if (!tDecodeIsEnd(&decoder)) {
-    TAOS_CHECK_EXIT(tDecodeU64(&decoder, &pReq->clientId));
-  } else {
-    pReq->clientId = 0;
+    pReq->pOpParam = taosMemoryCalloc(1, sizeof(SOperatorParam));
+    if (NULL == pReq->pOpParam) {
+      TAOS_CHECK_EXIT(terrno);
+    }
+    TAOS_CHECK_EXIT(tDeserializeSOperatorNotifyParam(&decoder, pReq->pOpParam));
  }


source/libs/planner/src/planSpliter.c

+      // TODO(tony): for debug!!!!!!!!!!!
+      // QUERY_NODE_LOGIC_PLAN_INTERP_FUNC != nodeType(pNode->pParent) &&


source/dnode/vnode/src/tsdb/tsdbRead2.c

+/*
+  @brief Mark the current step done, so next step will be triggered.
+  @param pReader the reader to mark the step done
+  @param notifyTs the timestamp to notify, used to determine whether  


source/dnode/vnode/src/tsdb/tsdbRead2.c

+  int32_t code = tsdbAcquireReader(pReader);
+  TSDB_CHECK_CODE(code, lino, _end);


source/libs/executor/src/executorInt.c

+  SOperatorParam* pParam = taosMemoryCalloc(1, sizeof(SOperatorParam));
+  QUERY_CHECK_NULL(pParam, code, lino, _end, terrno);
+  pParam->opType = opType;
+  pParam->value = taosMemoryCalloc(1, sizeof(int64_t));
+  QUERY_CHECK_NULL(pParam->value, code, lino, _end, terrno);


docs/zh/14-reference/03-taos-sql/20-select.md

  - RANGE: 指定单个或者开始结束时间值，结束时间须大于开始时间，ts_val 为标准时间戳类型，surrounding_time_val 可选，指定时间范围，为正值，精度可选 1n、1u、1a、1s、1m、1h、1d、1w。如 ```RANGE('2023-10-01T00:00:00.000')``` 、```RANGE('2023-10-01T00:00:00.000', '2023-10-01T23:59:59.999')```、```RANGE('2023-10-01T00:00:00.000', '2023-10-01T23:59:59.999'，1h)```。
  - EVERY: 时间间隔范围，every_val 为正值，精度可选 1n、1u、1a、1s、1m、1h、1d、1w，如 EVERY(1s)。
-  - FILL: 类型可选 NONE (不填充)、VALUE（指定值填充）、PREV（前一个非 NULL 值）、NEXT（后一个非 NULL）、NEAR（前后最近的非 NULL 值）。
+  - FILL: 类型可选 NONE (不填充)、VALUE（指定值填充）、PREV（前一个有效值）、NEXT（后一个有效值）、NEAR（最近邻的有效值）、LINEAR（线性插值）；注意，NULL值能否被认定为有效数据取决于 interp 函数的 ignore_null_values 参数。


docs/zh/14-reference/03-taos-sql/22-function.md

 ```

-**功能说明**：返回指定时间截面指定列的记录值或插值。ignore_null_values 参数的值可以是 0 或 1，为 1 时表示忽略 NULL 值，缺省值为 0。
+**功能说明**：返回指定时间截面指定列的记录值或插值。ignore_null_values 参数的值可以是 0 或 1，为 1 时表示忽略 NULL 值，缺省值为 0。当 ignore_null_values 为1时，插值时将会忽略其他 NULL 值采样数据。


dapan1121 · 2025-12-25T03:44:47Z

docs/zh/14-reference/03-taos-sql/24-distinguished.md

 1. 不进行填充：NONE（默认填充模式）。
 2. VALUE 填充：固定值填充，此时需要指定填充的数值。例如 `FILL(VALUE, 1.23)`。这里需要注意，最终填充的值受由相应列的类型决定，如 `FILL(VALUE, 1.23)`，相应列为 INT 类型，则填充值为 1，若查询列表中有多列需要 FILL，则需要给每一个 FILL 列指定 VALUE，如 `SELECT _wstart, min(c1), max(c1) FROM ... FILL(VALUE, 0, 0)`，注意，SELECT 表达式中只有包含普通列时才需要指定 FILL VALUE，如 `_wstart`、`_wstart+1a`、`now`、`1+1` 以及使用 `partition by` 时的 `partition key` (如 tbname) 都不需要指定 VALUE，如 `timediff(last(ts), _wstart)` 则需要指定 VALUE。
-3. PREV 填充：使用前一个值填充数据。例如 FILL(PREV)。
+3. PREV 填充：使用前一个有效值填充数据。例如 FILL(PREV)。


在这里把有效值的含义也分别说一下吧

dapan1121 · 2025-12-25T05:44:46Z

source/libs/executor/src/timesliceoperator.c

-    cond->type = TIMEWINDOW_RANGE_EXTERNAL;
-    code = getQueryExtWindow(&cond->twindows, &pInfo->win, &cond->twindows, cond->extTwindows);
-    QUERY_CHECK_CODE(code, lino, _error);
+    setOperatorNotifyFn(downstream, scanOptrNotifyReaderStepDone);


这个应该在各个算子里自己设置自己的吧

notify方法可能会受上游算子的影响，不同的上游算子可能notify的内容不一样，所以在上游的interp算子里设置了。

但是tablescan算子里自己设置了，因为qnode和vnode分离的时候，interp算子无法为tablescan算子设置这个接口，以后如果还有其他的notify函数的话这会是个问题。

dapan1121 · 2025-12-25T05:47:07Z

source/libs/executor/src/exchangeoperator.c

+           " since:%s", __func__, lino, sourceIdx, tstrerror(code));
+  }
+  taosMemoryFree(msg);
+  taosMemoryFree(pMsgSendInfo);


需要确认这种用法有没有问题

有问题，已经修改为noResp形式的rpcMsg接口

dapan1121 · 2025-12-25T05:53:01Z

source/dnode/vnode/src/tsdb/tsdbRead2.c

+  @param notifyTs the timestamp to notify, used to determine whether  
+         to mark the current step as done
+*/
+int32_t tsdbReaderStepDone(STsdbReader* pReader, int64_t notifyTs) {


这个函数跟scan算子有可能在两个线程上执行，需要有并发访问控制

通过pReader内部的互斥锁控制query和notify在不同线程并发访问

在释放taskHandle和pReader前给taskCtx上写锁来避免use-after-free error

…mprove query structure

…cases for filling nulls

…ior details

…n and optimize memory allocation

…for INTERVAL and INTERP clauses

This reverts commit 1a9c638.

Copilot AI review requested due to automatic review settings December 22, 2025 07:15

Tony2h requested review from a team, dapan1121, guanshengliang and hzcheng as code owners December 22, 2025 07:15

taosdata-bot bot added the team engine label Dec 22, 2025

Copilot started reviewing on behalf of Tony2h December 22, 2025 07:15 View session

gemini-code-assist bot reviewed Dec 22, 2025

View reviewed changes

Copilot AI reviewed Dec 22, 2025

View reviewed changes

gemini-code-assist bot reviewed Dec 22, 2025

View reviewed changes

Tony2h force-pushed the enh/3.0/fillinterp branch from 2ba83ee to 79c81c1 Compare December 23, 2025 10:05

Tony2h requested a review from zitsen as a code owner December 24, 2025 02:36

github-actions bot reviewed Dec 24, 2025

View reviewed changes

Tony2h force-pushed the enh/3.0/fillinterp branch 2 times, most recently from 2e97edd to 85862ef Compare December 24, 2025 08:49

dapan1121 reviewed Dec 25, 2025

View reviewed changes

Tony Zhang added 12 commits December 26, 2025 13:31

enh(interp): fill(prev/next/linear) can be confiured by ignoreNull param

84aa5ab

fix(interp): prev row escape error

05e135e

feat: imple exchange optr and executor to handle stepDone msg

24fab6c

impl notifyParam for stepDone notify msg

0959877

chore: fix format

b1d9dbb

refactor(interp): update test cases to include new fill options and i…

ffbe5ad

…mprove query structure

test: impl interp fill stream tests

351af85

refactor(interp): enhance query interpolation tests and add new test …

4b4722e

…cases for filling nulls

docs: clarify FILL clause descriptions and update interpolation behav…

dcdc12f

…ior details

refactor(notify): remove child parameter serialization/deserializatio…

771a823

…n and optimize memory allocation

fix: destroy node error

f4d09fb

fix: merge compile error

93088fe

Tony Zhang added 3 commits December 26, 2025 14:04

docs: enhance FILL clause documentation with details on valid values …

4321c9f

…for INTERVAL and INTERP clauses

refactor(notify): exchange optr send noresp rpcmsg to downstream

9bb69ae

refactor(reader): make reader thread safe

1a9c638

Tony2h force-pushed the enh/3.0/fillinterp branch from 43f250a to 1a9c638 Compare December 26, 2025 06:05

Tony Zhang added 5 commits December 26, 2025 14:12

chore: fix doc format

20c894f

Revert "refactor(reader): make reader thread safe"

6afe822

This reverts commit 1a9c638.

fix: reader use-after-free error

6d1eaf7

chore: print format

169878c

fix: dead lock on error path

23b53d7

		// TODO(tony): for debug!!!!!!!!!!!
		// QUERY_NODE_LOGIC_PLAN_INTERP_FUNC != nodeType(pNode->pParent) &&

		int32_t code = tsdbAcquireReader(pReader);
		TSDB_CHECK_CODE(code, lino, _end);

enh(interp): fill(prev/next/near/linear) is configured by ignoreNull param #34006

Are you sure you want to change the base?

enh(interp): fill(prev/next/near/linear) is configured by ignoreNull param #34006

Conversation

Tony2h commented Dec 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Issue(s)

Checklist

Uh oh!

Tony2h commented Dec 22, 2025

Uh oh!

gemini-code-assist bot commented Dec 22, 2025

Summary of Changes

Highlights

Footnotes

Uh oh!

gemini-code-assist bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

This comment was marked as spam.

Uh oh!

This comment was marked as abuse.

Uh oh!

This comment was marked as abuse.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as abuse.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as spam.

Uh oh!

This comment was marked as outdated.

Uh oh!

gemini-code-assist bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

This comment was marked as spam.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as duplicate.

Uh oh!

This comment was marked as duplicate.

Uh oh!

This comment was marked as spam.

Uh oh!

This comment was marked as spam.

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

Tony2h commented Dec 22, 2025 •

edited

Loading

Tony2h Dec 25, 2025 •

edited

Loading