TDengine provides many connectors for development, including C/C++, JAVA, Python, RESTful, Go, Node.js, etc.


At present, TDengine connectors support a wide range of platforms, including hardware platforms such as X64/X86/ARM64/ARM32/MIPS/Alpha, and development environments such as Linux/Win64/Win32. The comparison matrix is as follows:

CPU X64 64bit X64 64bit X64 64bit X86 32bit ARM64 ARM32 MIPS Godson Alpha Sunway X64 TimecomTech
OS Linux Win64 Win32 Win32 Linux Linux Linux Linux Linux
Python --
Go -- --
Node.js -- --
C# -- --

Note: ● stands for that has been verified by official tests; ○ stands for that has been verified by unofficial tests.


  • To access the TDengine database through connectors (except RESTful) in the system without TDengine server software, it is necessary to install the corresponding version of the client installation package to make the application driver (the file name is in Linux system and taos.dll in Windows system) installed in the system, otherwise, the error that the corresponding library file cannot be found will occur.
  • All APIs that execute SQL statements, such as taos_query(), taos_query_a(), taos_subscribe() in C/C++ connector, and APIs corresponding to them in other languages, can only execute one SQL statement at a time. If the actual parameters contain multiple statements, their behavior is undefined.
  • Users upgrading to TDengine 2.0. 8.0 must update the JDBC connection. TDengine must upgrade taos-jdbcdriver to 2.0.12 and above.
  • No matter which programming language connector is selected, TDengine version 2.0 and above recommends that each thread of database application establish an independent connection or establish a connection pool based on threads to avoid mutual interference between threads of "USE statement" state variables in the connection (but query and write operations of the connection are thread-safe).

Steps of Connector Driver Installation

The server should already have the TDengine server package installed. The connector driver installation steps are as follows:


1. Download from TAOS Data official website

  • X64 hardware environment: TDengine-client-2.x.x.x-Linux-x64.tar.gz
  • ARM64 hardware environment: TDengine-client-2.x.x.x-Linux-aarch64.tar.gz
  • ARM32 hardware environment: TDengine-client-2.x.x.x-Linux-aarch32.tar.gz

2. Unzip the package

Place the package in any directory that current user can read/write, and then execute following command:

tar -xzvf TDengine-client-xxxxxxxxx.tar.gz

Where xxxxxx needs to be replaced with you actual version as a string.

3. Execute installation script

After extracting the package, you will see the following files (directories) in the extracting directory:

install_client. sh: Installation script for application driver

taos.tar.gz: Application driver installation package

driver: TDengine application driver

connector: Connectors for various programming languages (Go/grafanaplugin/Node.js/Python/JDBC)

Examples: Sample programs for various programming languages (C/C#/Go/JDBC/MATLAB/Python/R)

Run to install.

4. Configure taos.cfg

Edit the taos.cfg file (default path/etc/taos/taos.cfg) and change firstEP to End Point of the TDengine server, for example:


1. If no TDengine service deployed in this machine, but only the application driver is installed, only firstEP needs to be configured in taos.cfg, and FQDN does not.

2. To prevent “unable to resolve FQDN” error when connecting to the server, ensure that the hosts file of the client has the correct FQDN value.

Windows x64/x86

1. Download from TAOS Data official website

  • X64 hardware environment: TDengine-client-2.X.X.X-Windows-x64.exe
  • X86 hardware environment: TDengine-client-2.X.X.X-Windows-x86.exe

2. Execute installation, select default values as prompted to complete

3. Installation path

Default installation path is: C:\TDengine, with following files(directories):

taos.exe: taos shell command line program

cfg: configuration file directory

driver: application driver dynamic link library

examples: sample program bash/C/C #/go/JDBC/Python/Node.js

include: header file

log: log file

unins000. exe: uninstall program

4. Configure taos.cfg

Edit the taos.cfg file (default path/etc/taos/taos.cfg) and change firstEP to End Point of the TDengine server, for example:


1. If you use FQDN to connect to the server, you must confirm that the DNS of the local network environment has been configured, or add FQDN addressing records in the hosts file. For example, edit C:\ Windows\ system32\ drivers\ etc\ hosts, and add the following record: 192.168. 1.99

2. Uninstall: Run unins000. exe to uninstall the TDengine application driver.

Installation verification

After the above installation and configuration completed, and confirm that the TDengine service has started running normally, the taos client can be logged in at this time.

Linux environment:

If you execute taos directly under Linux shell, you should be able to connect to TDengine service normally and jump to taos shell interface. For Example:

$ taos
Welcome to the TDengine shell from Linux, Client  Version:
Copyright (c) 2017 by TAOS Data, Inc. All rights  reserved.
taos> show databases;
name       |   created_time    |   ntables  |  vgroups   | replica | quorum | days |    keep1,keep2,keep(D)   | cache(MB)|   blocks  |  minrows   |  maxrows  | wallevel |  fsync    | comp | precision |    status  |
test       | 2020-10-14  10:35:48.617 |     10 |      1 |    1 |   1 |     2 | 3650,3650,3650        |     16|      6 |     100 |    4096 |    1 |    3000 |  2 | ms      | ready    |
log        | 2020-10-12  09:08:21.651 |      4 |      1 |    1 |   1 |   10 | 30,30,30               |      1|      3 |     100 |    4096 |    1 |    3000 |  2 | us    | ready    |
Query OK, 2 row(s) in set (0.001198s)

Windows (x64/x86) environment:

Under cmd, enter the c:\TDengine directory and directly execute taos.exe, and you should be able to connect to TDengine service normally and jump to taos shell interface. For example:

  Welcome to the TDengine  shell from Linux, Client Version:
  Copyright (c) 2017 by  TAOS Data, Inc. All rights reserved.
  taos> show  databases;
  name       |   created_time    |   ntables  |  vgroups   | replica | quorum | days |    keep1,keep2,keep(D)   | cache(MB)   |  blocks  |   minrows  |  maxrows   | wallevel |  fsync  | comp | precision |  status    |
  test       | 2020-10-14  10:35:48.617 |     10 |      1 |    1 |   1 |     2 | 3650,3650,3650        |     16 |      6 |     100 |    4096 |    1 |    3000 |  2 | ms    | ready    |
  log        | 2020-10-12  09:08:21.651 |      4 |      1 |    1 |   1 |    10 | 30,30,30              |      1 |      3 |     100 |    4096 |    1 |    3000 |  2 | us    | ready    |
  Query OK, 2 row(s) in  set (0.045000s)

C/C++ Connector

Systems supported by C/C++ connectors as follows:

CPU Type x64**64bit)** ARM64 ARM32
OS Type Linux Win64 Win32 Linux Linux
Supported or Not Yes Yes Yes Yes Yes

The C/C++ API is similar to MySQL's C API. When application use it, it needs to include the TDengine header file taos.h (after installed, it is located in/usr/local/taos/include):

#include <taos.h>


  • The TDengine dynamic library needs to be linked at compiling. The library in Linux is, which installed at/usr/local/taos/driver. By Windows, it is taos.dll and installed at C:\ TDengine.
  • Unless otherwise specified, when the return value of API is an integer, 0 represents success, others are error codes representing the cause of failure, and when the return value is a pointer, NULL represents failure.

More sample codes for using C/C++ connectors, please visit

Basic API

The basic API is used to create database connections and provide a runtime environment for the execution of other APIs.

  • void taos_init()

Initialize the running environment. If the application does not actively call the API, the API will be automatically called when the application call taos_connect(), so the application generally does not need to call the API manually.

  • void taos_cleanup()

Clean up the running environment and call this API before the application exits.

  • int taos_options(TSDB_OPTION option, const void * arg, ...)

Set client options, currently only time zone setting (_TSDB_OPTIONTIMEZONE) and encoding setting (_TSDB_OPTIONLOCALE) are supported. The time zone and encoding default to the current operating system settings.

When the return value is 0, it means success, and when it is -1, it means failure.

  • char *taos_get_client_info()

Get version information of the client.

  • TAOS *taos_connect(const char *host, const char *user, const char *pass, const char *db, int port)

Create a database connection and initialize the connection context. The parameters that need to be provided by user include:

  • host: FQDN used by TDengine to manage the master node
  • user: User name
  • pass: Password
  • db: Database name. If user does not provide it, it can be connected normally, means user can create a new database through this connection. If user provides a database name, means the user has created the database and the database is used by default
  • port: Port number

A null return value indicates a failure. The application needs to save the returned parameters for subsequent API calls. Note: The same process can connect to multiple taosd processes based on ip/port

  • char *taos_get_server_info(TAOS *taos)

Get version information of the server-side.

  • int taos_select_db(TAOS *taos, const char *db)

Set the current default database to db.

The return value is the error code.

  • void taos_close(TAOS *taos)

Close the connection, where taos is the pointer returned by taos_connect() function.

Synchronous query API

Traditional database operation APIs all make synchronous operations. After the application calls an API, it remains blocked until the server returns the result. TDengine supports the following APIs:

  • TAOS_RES* taos_query(TAOS *taos, const char *sql)

This API is used to execute SQL statements, which can be DQL, DML or DDL statements. Where taos parameter is a pointer obtained through taos_connect(). You can't judge whether the execution result fails by whether the return value is NULL, but to use taos_errno() function to parse the error code in the result set.

  • int taos_result_precision(TAOS_RES *res)

The precision of the timestamp field in the returned result set, 0 for milliseconds, 1 for microseconds, and 2 for nanoseconds.

  • TAOS_ROW taos_fetch_row(TAOS_RES *res)

Get the data in the query result set by rows.

  • int taos_fetch_block(TAOS_RES *res, TAOS_ROW *rows)

The data in the query result set is obtained in batch, and the return value is the number of rows of the obtained data.

  • int taos_num_fields(TAOS_RES *res)int taos_field_count(TAOS_RES *res)

The two APIs are equivalent, and are used to get the number of columns in the query result set.

  • int* taos_fetch_lengths(TAOS_RES *res)

Get the length of each field in the result set. The return value is an array whose length is the number of columns in the result set.

  • int taos_affected_rows(TAOS_RES *res)

Get the number of rows affected by the executed SQL statement.

  • TAOS_FIELD *taos_fetch_fields(TAOS_RES *res)

Get the attributes (data type, name, number of bytes) of each column of data in the query result set, which can be used in conjunction with taos_num_files to parse the data of a tuple (one row) returned by taos_fetch_row(). The structure of TAOS_FIELD is as follows:

typedef struct taosField {
  char     name[65];  // Column name
  uint8_t  type;      // Data type
  int16_t  bytes;     // Number of bytes
  • void taos_stop_query(TAOS_RES *res)

Stop the execution of a query.

  • void taos_free_result(TAOS_RES *res)

Release the query result set and related resources. After the query is completed, be sure to call the API to release resources, otherwise it may lead to application memory leakage. However, it should also be noted that after releasing resources, if you call functions such as taos_consume() to obtain query results, it will lead the application to Crash.

  • char *taos_errstr(TAOS_RES *res)

Get the reason why the last API call failed, and the return value is a string.

  • char *taos_errno(TAOS_RES *res)

Get the reason why the last API call failed, and the return value is the error code.

Note: TDengine 2.0 and above recommends that each thread of a database application establish an independent connection or establish a connection pool based on threads. It is not recommended to pass the connection (TAOS*) structure to different threads for sharing in applications. Query and write operations based on TAOS structure have multithread safety, but state variables such as "USE statement" may interfere with each other among threads. In addition, C connector can dynamically establish new database-oriented connections according to requirements (this process is not visible to users), and it is recommended to call taos_close() to close the connection only when the program finally exits.

Asynchronous query API

In addition to synchronous API, TDengine also provides higher performance asynchronous call API to handle data insertion and query operations. Under the same software and hardware environment, asynchronous API processes data insertion 2 ~ 4 times faster than synchronous API. Asynchronous API adopts a non-blocking call mode and returns immediately before the system really completes a given database operation. The calling thread can handle other work, thus improving the performance of the whole application. Asynchronous API has outstanding advantages in the case of poor network delay.

Asynchronous APIs all need applications to provide corresponding callback function. The callback function parameters are set as follows: the first two parameters are consistent, and the third parameter depends on different APIs. The first parameter param is provided to the system when the application calls the asynchronous API. When used for callback, the application can retrieve the context of the specific operation, depending on the specific implementation. The second parameter is the result set of SQL operation. If it is empty, such as insert operation, it means that there is no record returned. If it is not empty, such as select operation, it means that there is record returned.

Asynchronous APIs have relatively high requirements for users, who can selectively use them according to specific application scenarios. Here are three important asynchronous APIs:

  • void taos_query_a(TAOS *taos, const char *sql, void (*fp)(void *param, TAOS_RES *, int code), void *param); Execute SQL statement asynchronously.

    • taos: The database connection returned by calling taos_connect()
    • sql: The SQL statement needed to execute
    • fp: User-defined callback function, whose third parameter code is used to indicate whether the operation is successful, 0 for success, and negative number for failure (call taos_errstr() to get the reason for failure). When defining the callback function, it mainly handles the second parameter TAOS_RES *, which is the result set returned by the query
    • param:the parameter for the callback
  • void taos_fetch_rows_a(TAOS_RES *res, void (*fp)(void *param, TAOS_RES *, int numOfRows), void *param); Get the result set of asynchronous queries in batch, which can only be used with taos_query_a(). Within:

    • res: The result set returned when taos_query_a() callback.
    • fp: Callback function. Its parameter param is a user-definable parameter construct passed to the callback function; numOfRows is the number of rows of data obtained (not a function of the entire query result set). In the callback function, applications can get each row of the batch records by calling taos_fetch_rows() forward iteration. After reading all the records in a block, the application needs to continue calling taos_fetch_rows_a() in the callback function to obtain the next batch of records for processing until the number of records returned (numOfRows) is zero (the result is returned) or the number of records is negative (the query fails).

The asynchronous APIs of TDengine all use non-blocking calling mode. Applications can use multithreading to open multiple tables at the same time, and can query or insert to each open table at the same time. It should be pointed out that the application client must ensure that the operation on the same table is completely serialized, that is, when the insertion or query operation on the same table is not completed (when no result returned), the second insertion or query operation cannot be performed.

Parameter binding API

In addition to calling taos_query() directly for queries, TDengine also provides a Prepare API that supports parameter binding. Like MySQL, these APIs currently only support using question mark ? to represent the parameters to be bound, as follows:

  • TAOS_STMT* taos_stmt_init(TAOS *taos)

Create a TAOS_STMT object for calling later.

  • int taos_stmt_prepare(TAOS_STMT *stmt, const char *sql, unsigned long length)

Parse a SQL statement and bind the parsing result and parameter information to STMT. If the parameter length is greater than 0, this parameter will be used as the length of the SQL statement. If it is equal to 0, the length of the SQL statement will be automatically judged.

  • int taos_stmt_bind_param(TAOS_STMT *stmt, TAOS_BIND *bind)

For parameter binding, bind points to an array, and it is necessary to ensure that the number and order of elements in this array are exactly the same as the parameters in sql statement. TAOS_BIND is used in the same way as MYSQL_BIND in MySQL and is defined as follows:

typedef struct TAOS_BIND {
  int            buffer_type;
  void *         buffer;
  unsigned long  buffer_length;  // Not in use
  unsigned long *length;
  int *          is_null;
  int            is_unsigned;    // Not in use
  int *          error;          // Not in use

Add the current bound parameters to the batch. After calling this function, you can call taos_stmt_bind_param() again to bind the new parameters. It should be noted that this function only supports insert/import statements, and if it is other SQL statements such as select, it will return errors.

  • int taos_stmt_execute(TAOS_STMT *stmt)

Execute the prepared statement. At the moment, a statement can only be executed once.

  • TAOS_RES* taos_stmt_use_result(TAOS_STMT *stmt)

Gets the result set of the statement. The result set is used in the same way as when calling non-parameterized. After using it, taos_free_result() should be called to release resources.

  • int taos_stmt_close(TAOS_STMT *stmt)

Execution completed, release all resources.

  • char * taos_stmt_errstr(TAOS_STMT *stmt)

Gets the error message if any stmt API returns error.

Schemaless writing API

In addition to writing data using SQL or using the parameter binding API, writing can also be done using Schemaless, which eliminates the need to create a super table/data sub-table data structure in advance and writes data directly, while the TDengine system automatically creates and maintains the required table structure based on the written data content. The use of Schemaless is described in the Schemaless Writing section, and the C/C++ API used with it is described here.

  • TAOS_RES* taos_schemaless_insert(TAOS* taos, const char* lines[], int numLines, int protocol, int precision)

    Function Description

    This interface writes the text data of the line protocol to TDengine.

    Parameter Description

    taos: database connection, the database connection established by taos_connect function.

    lines: text data. A pattern-free text string that meets the parsing format requirements.

    numLines: the number of lines of the text data, cannot be 0.

    protocol: the protocol type of the lines, used to identify the format of the text data.

    precision: precision string of the timestamp in the text data.

    Return Value

    TAOS_RES structure, the application can get the error message by using taos_errstr and also get the error code by using taos_errno.

    In some cases, the returned TAOS_RES is NULL, in which case taos_errno can still be called to safely get the error code information.

    The returned TAOS_RES needs to be freed by the caller, otherwise a memory leak will occur.


    The protocol type is enumerated and contains the following three formats.

    TSDB_SML_LINE_PROTOCOL: InfluxDB line protocol (Line Protocol)


    TSDB_SML_JSON_PROTOCOL: OpenTSDB JSON protocol format

    The timestamp resolution is defined in the taos.h file, as follows








    Note that the timestamp resolution parameter only takes effect when the protocol type is SML_LINE_PROTOCOL.

    For OpenTSDB text protocols, the timestamp resolution follows the official resolution rules - the time precision is determined by the number of characters contained in the timestamp.

    Supported versions

    This functional interface is supported since version

#include <stdlib.h>
#include <stdio.h>
#include <taos.h>

int main() {
  const char* host = "";
  const char* user = "root";
  const char* passwd = "taosdata";

  // connect to server
  TAOS* taos = taos_connect(host, user, passwd, "test", 0);

  // prepare the line string
  char* lines1[] = {
      "stg,t1=3i64,t2=4f64,t3=\"t3\" c1=3i64,c3=L\"passit\",c2=false,c4=4f64 1626006833639000000",
      "stg,t1=4i64,t3=\"t4\",t2=5f64,t4=5f64 c1=3i64,c3=L\"passitagin\",c2=true,c4=5f64,c5=5f64 1626006833641000000"

  // schema-less insert
  TAOS_RES* res = taos_schemaless_insert(taos, lines1, 2, TSDB_SML_LINE_PROTOCOL, TSDB_SML_TIMESTAMP_NANO_SECONDS);
  if (taos_errno(res) != 0) {
    printf("failed to insert schema-less data, reason: %s\n", taos_errstr(res));


  // close the connection
  return (code);

Data subscription interface

The subscription API currently supports subscribing to one or more tables and continuously obtaining the latest data written to the tables through regular polling.

  • TAOS_SUB *taos_subscribe(TAOS* taos, int restart, const char* topic, const char *sql, TAOS_SUBSCRIBE_CALLBACK fp, void *param, int interval)

This function is for starting the subscription service, returning the subscription object in case of success, and NULL in case of failure. Its parameters are:

  • taos: Database connection established
  • Restart: If the subscription already exists, do you want to start over or continue with the previous subscription
  • Topic: Subject (that is, name) of the subscription. This parameter is the unique identification of the subscription
  • sql: The query statement subscribed. This statement can only be a select statement. It should only query the original data, and can only query the data in positive time sequence
  • fp: The callback function when the query result is received (the function prototype will be introduced later). It is only used when calling asynchronously, and this parameter should be passed to NULL when calling synchronously
  • param: The additional parameter when calling the callback function, which is passed to the callback function as it is by the system API without any processing
  • interval: Polling period in milliseconds. During asynchronous call, the callback function will be called periodically according to this parameter; In order to avoid affecting system performance, it is not recommended to set this parameter too small; When calling synchronously, if the interval between two calls to taos_consume() is less than this period, the API will block until the interval exceeds this period.
  • typedef void (*TAOS_SUBSCRIBE_CALLBACK)(TAOS_SUB* tsub, TAOS_RES *res, void* param, int code)

In asynchronous mode, the prototype of the callback function has the following parameters:

  • tsub: Subscription object
  • res: Query the result set. Note that there may be no records in the result set
  • param: Additional parameters supplied by the client when taos_subscribe() is called
  • code: Error code
  • TAOS_RES *taos_consume(TAOS_SUB *tsub)

In synchronous mode, this function is used to get the results of subscription. The user application places it in a loop. If the interval between two calls to taos_consume() is less than the polling cycle of the subscription, the API will block until the interval exceeds this cycle. If a new record arrives in the database, the API will return the latest record, otherwise it will return an empty result set with no records. If the return value is NULL, it indicates a system error. In asynchronous mode, user program should not call this API.

  • void taos_unsubscribe(TAOS_SUB *tsub, int keepProgress)

Unsubscribe. If the parameter keepProgress is not 0, the API will keep the progress information of subscription, and the subsequent call to taos_subscribe() can continue based on this progress; otherwise, the progress information will be deleted and the data can only be read again.

Python Connector

See video tutorials for the use of Python connectors.

Installation preparation

Python connector installation

From TDengine 2.4, users can install python connector for TDengine with pip. Note that the package name is taospy (not taos - a fully unrelated package). For backward compatibility, we still use import taos to import connector package.

pip install taospy

Use your version-specific pip command as if you need.

pip2 install taospy
pip3 install taospy

The python connector requires libtaos library ( in Linux, or taos.dll in Windows). For Windows client, if import taos failed, you could copy the dll C:\TDengine\driver\taos.dll to C:\windows\system32 and try it again.

For users that has a limited network environment, just add the connector/python of installed directory(commonly /usr/local/taos/connector/python/ in Linux, C:\TDengine\connector\python in Windows) to PYTHONPATH environment variable.

How to use

PEP-249 Python Database API

Definitely you can use the PEP-249 database API like other type of databases:

import taos

conn = taos.connect()
cursor = conn.cursor()

cursor.execute("show databases")
results = cursor.fetchall()
for row in results:
Code sample
  • Import the TDengine client module

    import taos
  • Get the connection and cursor object

    conn = taos.connect(host="", user="root", password="taosdata", config="/etc/taos")
    c1 = conn.cursor()

    host covers all IPs of TDengine server-side, and config is the directory where the client configuration files is located

  • Write data

    import taos
    import datetime
    conn = taos.connect()
    c1 = conn.cursor()
    # Create a database
    c1.execute('create database if not exists db1')
    c1.execute('use db1')
    # Create a table
    c1.execute('create table if not exists tb (ts timestamp, temperature int, humidity float)')
    # Insert data
    start_time = datetime.datetime(2019, 11, 1)
    affected_rows = c1.execute('insert into tb values (\'%s\', 0, 0.0)' %start_time)
    # Insert data in batch
    time_interval = datetime.timedelta(seconds=60)
    sqlcmd = ['insert into tb values']
    for irow in range(1,11):
        start_time += time_interval
        sqlcmd.append('(\'%s\', %d, %f)' %(start_time, irow, irow*1.2))
    affected_rows += c1.execute(' '.join(sqlcmd))
    print("inserted %s records" % affected_rows)
  • Query data

    c1.execute('select * from tb')
    # pull query result
    data = c1.fetchall()
    # The result is a list, with each row as an element
    numOfRows = c1.rowcount
    numOfCols = len(c1.description)
    for irow in range(numOfRows):
        print("Row%d: ts=%s, temperature=%d, humidity=%f" %(irow, data[irow][0], data[irow][1],data[irow][2]))
    # Use cursor loop directly to pull query result
    c1.execute('select * from tb')
    for data in c1:
        print("ts=%s, temperature=%d, humidity=%f" %(data[0], data[1],data[2]))

Query API

Since v2.1.0, python connector provides a new API for query:

import taos

conn = taos.connect()
conn.execute("create database if not exists pytest")

result = conn.query("show databases")
num_of_fields = result.field_count
for field in result.fields:
for row in result:
conn.execute("drop database pytest")

The query method returns TaosResult class. It provides high level APIs for convenient use:


  • fields: the TaosFields object contains the column metadata, given the collection of each column field metadata by iterator.
  • field_count: column number of result.
  • affected_rows: the rows completed for insert.
  • row_count: the rows number for select.
  • precision: the result precision.


  • fetch_all(): get all data as tuple array.
  • fetch_all_into_dict(): get all data as dict array, added since v2.1.1
  • blocks_iter(): provides iterator by C taos_fetch_blocks() API
  • rows_iter(): provides iterator by C taos_fetch_row() API
  • fetch_rows_a: fetch rows by async API in taosc.
  • errno: error code if failed.
  • errstr: error string if failed.
  • close: close result, you do not need to call it directly, result will auto closed out of scope.

Subscription API

Create subscription

# Create a subscription with the topic ‘test’ and a consumption cycle of 1000 milliseconds
# If the first parameter is True, it means restarting the subscription.
# If it is False and a subscription with the topic 'test 'has been created before,
# it means continuing to consume the data of this subscription instead of restarting to consume all the data
sub = conn.subscribe(True, "test", "select * from tb;", 1000)

Consume subscription data.

data = sub.consume()
for d in data:



Close connection.


JSON Type Support

Python connector taospy starts supporting JSON type as tags since v2.2.0 (requires TDengine beta v2.3.5+, or stable v2.4.0+).

Create stable and table with JSON tag.

# encoding:UTF-8
import taos

conn = taos.connect()
conn.execute("create database if not exists py_test_json_type")
conn.execute("use py_test_json_type")

conn.execute("create stable s1 (ts timestamp, v1 int) tags (info json)")
conn.execute("create table s1_1 using s1 tags ('{\"k1\": \"v1\"}')")

Query JSON tag and table name from a stable.

tags = conn.query("select info, tbname from s1").fetch_all_into_dict()

The tags value is:

[{'info': '{"k1":"v1"}', 'tbname': 's1_1'}]

To get value from JSON tag by key:

k1 = conn.query("select info->'k1' as k1 from s1").fetch_all_into_dict()
>>> k1
[{'k1': '"v1"'}]

Refer to JSON type instructions for more usage of JSON type.

Using nanosecond in Python connector

So far Python still does not completely support nanosecond type. Please refer to the link 1 and 2. The implementation of the python connector is to return an integer number for nanosecond value rather than datatime type as what ms and us do. The developer needs to handle it themselves. We recommend using pandas to_datetime() function. If Python officially support nanosecond in the future, TAOS Data might be possible to change the interface accordingly, which mean the application need change too.



Users can directly view the usage information of the module through Python's helper, or refer to the sample program in the examples/Python directory. The following are some common classes and methods:

  • TDengineConnection class

Refer to help (taos.TDengineConnection) in python. This class corresponds to a connection between the client and TDengine. In the scenario of client multithreading, it is recommended that each thread apply for an independent connection instance, but not recommended that multiple threads share a connection.

  • TDengineCursor class

Refer to help (taos.TDengineCursor) in python. This class corresponds to the write and query operations performed by the client. In the scenario of client multithreading, this cursor instance must be kept exclusive to threads and cannot be used by threads, otherwise errors will occur in the returned results.

  • connect method

Used to generate an instance of taos.TDengineConnection.

Python connector code sample

In the examples/python directory, we provide a sample Python program read_example. py to guide you to design your own write and query program. After installing the corresponding client, introduce the taos class through import taos. The steps are as follows:

  • Get the TDengineConnection object through taos.connect, which can be applied for only one by a program and shared among multiple threads.

  • Get a new cursor object through the .cursor () method of the TDengineConnection object, which must be guaranteed to be exclusive to each thread.

  • Execute SQL statements for writing or querying through the execute() method of the cursor object.

  • If a write statement is executed, execute() returns the number of rows successfully written affected rows.

  • If the query statement is executed, the result set needs to be pulled through the fetchall() method after the execution is successful.

    You can refer to the sample code for specific methods.

RESTful Connector

To support the development of various types of platforms, TDengine provides an API that conforms to REST design standards, that is, RESTful API. In order to minimize the learning cost, different from other designs of database RESTful APIs, TDengine directly requests SQL statements contained in BODY through HTTP POST to operate the database, and only needs a URL. See the video tutorial for the use of RESTful connectors.

Note: One difference from the native connector is that the RESTful interface is stateless, so the USE db_name command has no effect and all references to table names and super table names require the database name to be specified. (Starting from version, we support specifying db_name in the RESTful url, in which case if the database name prefix is not specified in the SQL statement. Since version, RESTful service is provided by taosAdapter by default, which requires that db_name must be specified in the url.)

HTTP request format


Parameter description:

  • IP: Any host in the cluster
  • PORT: httpPort configuration item in the configuration file, defaulting to 6041

For example: is a URL that points to an IP address of

The header of HTTP request needs to carry identity authentication information. TDengine supports Basic authentication and custom authentication. Subsequent versions will provide standard and secure digital signature mechanism for identity authentication.

  • Custom identity authentication information is as follows (We will introduce later)
Authorization: Taosd <TOKEN>
  • Basic identity authentication information is as follows
Authorization: Basic <TOKEN>

The BODY of HTTP request is a complete SQL statement. The data table in the SQL statement should provide a database prefix, such as \.\. If the table name does not have a database prefix, the system returns an error. Because the HTTP module is just a simple forwarding, there is no current DB concept.

Use curl to initiate an HTTP Request through custom authentication. The syntax is as follows:

curl -H 'Authorization: Basic <TOKEN>' -d '<SQL>' <ip>:<PORT>/rest/sql


curl -u username:password -d '<SQL>' <ip>:<PORT>/rest/sql

Where TOKEN is the string of {username}:{password} encoded by Base64, for example, root:taosdata will be encoded as cm9vdDp0YW9zZGF0YQ==.

HTTP return format

The return value is in JSON format, as follows:

    "status": "succ",
    "head": ["ts","current",...],
    "column_meta": [["ts",9,8],["current",6,4], ...],
    "data": [
        ["2018-10-03 14:38:05.000", 10.3, ...],
        ["2018-10-03 14:38:15.000", 12.6, ...]
    "rows": 2


  • status: Informs whether the operation results are successful or failed.
  • head: The definition of the table, with only one column "affected_rows" if no result set is returned. (Starting from version 2.0. 17, it is recommended not to rely on the head return value to judge the data column type, but to use column_meta. In future versions, head may be removed from the return value.)
  • column_meta: Starting with version 2.0. 17, this item is added to the return value to indicate the data type of each column in the data. Each column will be described by three values: column name, column type and type length. For example, ["current", 6, 4] means that the column name is "current"; the column type is 6, that is, float type; the type length is 4, which corresponds to a float represented by 4 bytes. If the column type is binary or nchar, the type length indicates the maximum content length that the column can save, rather than the specific data length in this return value. When the column type is nchar, its type length indicates the number of Unicode characters that can be saved, not bytes.
  • data: The specific returned data, rendered line by line, if no result set is returned, then only [[affected_rows]]. The order of the data columns for each row in data is exactly the same as the order of the data columns described in column_meta.
  • rows: Indicates the total number of rows of data.

Column types in column_meta:

  • 1:BOOL
  • 4:INT
  • 5:BIGINT
  • 6:FLOAT
  • 7:DOUBLE
  • 8:BINARY
  • 10:NCHAR

Custom authorization code

The HTTP request requires the authorization code <TOKEN> for identification. Authorization codes are usually provided by administrators. Authorization codes can be obtained simply by sending HTTP GET requests as follows:

curl http://<ip>:6041/rest/login/<username>/<password>

Where ip is the IP address of the TDengine database, username is the database user name, password is the database password, and the return value is in JSON format. The meanings of each field are as follows:

  • status: flag bit for request result
  • code: code of return value
  • desc: Authorization code

Sample to get authorization code:


Return value:

  "status": "succ",
  "code": 0,
  "desc": "/KfeAzX/f9na8qdtNZmtONryp201ma04bEl8LcvLUd7a8qdtNZmtONryp201ma04"

Use case

  • Lookup all records of table d1001 in demo database:
curl -H 'Authorization: Basic cm9vdDp0YW9zZGF0YQ==' -d 'select * from demo.d1001'

Return value:

    "status": "succ",
    "head": ["ts","current","voltage","phase"],
    "column_meta": [["ts",9,8],["current",6,4],["voltage",4,4],["phase",6,4]],
    "data": [
        ["2018-10-03 14:38:05.000",10.3,219,0.31],
        ["2018-10-03 14:38:15.000",12.6,218,0.33]
    "rows": 2
  • Create a database demo:
curl -H 'Authorization: Basic cm9vdDp0YW9zZGF0YQ==' -d 'create database demo'

Return value:

    "status": "succ",
    "head": ["affected_rows"],
    "column_meta": [["affected_rows",4,4]],
    "data": [[1]],
    "rows": 1

Other cases

Result set in Unix timestamp

When the HTTP request URL is sqlt, the timestamp of the returned result set will be expressed in Unix timestamp format, for example:

curl -H 'Authorization: Basic cm9vdDp0YW9zZGF0YQ==' -d 'select * from demo.d1001'

Return value:

    "status": "succ",
    "head": ["ts","current","voltage","phase"],
    "column_meta": [["ts",9,8],["current",6,4],["voltage",4,4],["phase",6,4]],
    "data": [
    "rows": 2

Result set in UTC time string

When the HTTP request URL is sqlutc, the timestamp of the returned result set will be represented by a UTC time string, for example:

  curl -H 'Authorization: Basic cm9vdDp0YW9zZGF0YQ==' -d 'select * from demo.t1'

Return value:

    "status": "succ",
    "head": ["ts","current","voltage","phase"],
    "column_meta": [["ts",9,8],["current",6,4],["voltage",4,4],["phase",6,4]],
    "data": [
    "rows": 2

Important configuration options

Only some configuration parameters related to RESTful interface are listed below. Please refer to the instructions in the configuration file for other system parameters. Note: After the configuration is modified, the taosd service needs to be restarted before it can take effect.

  • httpPort: The port number that provides RESTful services externally, which is bound to 6041 by default
  • httpMaxThreads: The number of threads started, the default is 2 (starting with version 2.0. 17, the default value is changed to half of the CPU cores and rounded down)
  • restfulRowLimit: The maximum number of result sets returned (in JSON format), default 10240
  • httpEnableCompress: Compression is not supported by default. Currently, TDengine only supports gzip compression format
  • httpdebugflag: Logging switch, 131: error and alarm information only, 135: debugging information, 143: very detailed debugging information, default 131

CSharp Connector

  • The C # connector supports: Linux 64/Windows x64/Windows x86.
  • C# connector can be download and include as normal table form
  • On Windows, C # applications can use the native C interface of TDengine to perform all database operations, and future versions will provide the ORM (Dapper) framework driver.

Installation preparation

  • For application driver installation, please refer to the steps of installing connector driver.
  • .NET interface file TDengineDrivercs.cs and reference sample TDengineTest.cs are both located in the Windows client install_directory/examples/C# directory.
  • Install .NET SDK

Example Source Code

you can find sample code under follow directions:

Tips: TDengineTest.cs One of C# connector's sample code that include basic examples like connection,sql executions and so on.

Installation verification

Run {client_install_directory}/examples/C#/C#Checker/C#Checker.cs

Need install .NET SDK first

cd {client_install_directory}/examples/C#/C#Checker
//run c#checker.cs
dotnet run -- -h <FQDN> //dotnet run will build project first by default.

How to use C# connector

On Windows system, .NET applications can use the .NET interface of TDengine to perform all database operations. The steps to use it are as follows:

need to install .NET SDK first

  • create a c# project.
mkdir test
cd test
dotnet new console
  • add TDengineDriver as an package through Nuget
dotnet add package TDengine.Connector
  • include the TDnengineDriver in you application's namespace
using TDengineDriver;
  • user can reference fromTDengineTest.cs and learn how to define database connection,query,insert and other basic data manipulations.


  • TDengine V2.0. 3.0 supports both 32-bit and 64-bit Windows systems, so when. NET project generates a .exe file, please select the corresponding "X86" or "x64" for the "Platform" under "Solution"/"Project".
  • This. NET interface has been verified in Visual Studio 2015/2017, and other VS versions have yet to be verified.
  • Since this. NET connector interface requires the taos.dll file, so before executing the application, copy the taos.dll file in the Windows {client_install_directory}/driver directory to the folder where the. NET project finally generated the .exe executable file. After running the exe file, you can access the TDengine database and do operations such as insert and query(This step can be skip if the client has been installed on you machine).

Third-party Driver

Maikebing.Data.Taos is an ADO.NET provider for TDengine that supports Linux, Windows. This development package is provided by enthusiastic contributor 麦壳饼@@maikebing. For more details:

// Download
// How to use

Go Connector

Installation preparation

The TDengine provides the GO driver taosSql. taosSql implements the GO language's built-in interface database/sql/driver. Users can access TDengine in the application by simply importing the package as follows, see for details.

Sample code for using the Go connector can be found in

import (
    _ ""

It is recommended to use Go version 1.13 or above and turn on module support:

go env -w GO111MODULE=on
go env -w GOPROXY=,direct

taosSql v2 completed refactoring of the v1 version and separated the built-in database operation interface database/sql/driver to the directory taosSql, and put other advanced functions such as subscription and stmt into the directory af.

Common APIs

  • sql.Open(DRIVER_NAME string, dataSourceName string) *DB

This API is used to open DB and return an object of type * DB. Generally, DRIVER_NAME is set to the string taosSql, and dataSourceName is set to the string user:password@/tcp(host:port)/dbname. If the customer wants to access TDengine with multiple goroutines concurrently, it is necessary to create a sql.Open object in each goroutine and use it to access TDengine.

Note: When the API is successfully created, there is no permission check. Only when Query or Exec is actually executed can the connection be truly created and whether the user/password/host/port is legal can be checked at the same time. In addition, because most of the implementation of the whole driver sinks into libtaos, which taosSql depends on. Therefore, sql.Open itself is particularly lightweight.

  • func (db *DB) Exec(query string, args ...interface{}) (Result, error)

sql.Open built-in method to execute non-query related SQL

  • func (db *DB) Query(query string, args ...interface{}) (*Rows, error)

sql.Open built-in method used to execute query statements

  • func (db *DB) Prepare(query string) (*Stmt, error)

sql.Open built-in method used to create a prepared statement for later queries or executions.

  • func (s *Stmt) Exec(args ...interface{}) (Result, error)

sql.Open built-in method to execute a prepared statement with the given arguments and returns a Result summarizing the effect of the statement.

  • func (s *Stmt) Query(args ...interface{}) (*Rows, error)

sql.Open built-in method to query executes a prepared query statement with the given arguments and returns the query results as a *Rows.

  • func (s *Stmt) Close() error

sql.Open built-in method to closes the statement.

Node.js Connector

The Node.js connector supports the following systems:

CPU Type x64(64bit) aarch64 aarch32
OS Type Linux Win64 Win32 Linux Linux
Supported or Not Yes Yes Yes Yes Yes

See the video tutorial for use of the Node.js connector.

Installation preparation

Install Node.js connector

Users can install it through npm or through the source code src/connector/nodejs/. The specific installation steps are as follows:

First, install the Node.js connector through npm.

npm install td2.0-connector

We recommend that use npm to install the Node.js connector. If you do not have npm installed, you can copy src/connector/nodejs/ to your Node.js project directory.

We use node-gyp to interact with the TDengine server. Before installing the Node.js connector, you also need to install the following software:


  • python (recommended v2.7, not currently supported in v3.x.x)
  • node 2.0. 6 supports v12. x and v10. x, 2.0. 5 and earlier supports v10. x, and other versions may have package compatibility issues.
  • make
  • GCC and other C compilers


Solution 1

Use Microsoft windows-build-tools to install all necessary tools by executing npm install --global --production windows-build-tools in cmd command line interface.

Solution 2

Manually install the following tools:

If the steps above cannot be performed successfully, you can refer to Microsoft's Node.js User Manual Microsoft's Node.js Guidelines for Windows.

If you use ARM64 Node.js on Windows 10 ARM, you also need to add "Visual C++ compilers and libraries for ARM64" and "Visual C++ ATL for ARM64".


The sample program source code is located in install_directory/examples/nodejs, and there are:

Node-example.js Node.js sample source code Node-example-raw.js

Installation verification

After installing the TDengine client, the nodejsChecker.js program can verify whether the current environment supports access to TDengine via Node.js.


  1. Create a new installation verification directory, for example: ~/tdengine-test, copy the nodejsChecker.js source program on GitHub. Download address:

  2. Execute the following command:

    npm init -y
    npm install td2.0-connector
    node nodejsChecker.js host=localhost
  3. After performing the above steps, the Node.js connection TDengine instance will be outputted on the command line, and the short-answer of insertion and query will be executed.

How to use Node.js

The following are some basic uses of Node.js connector. Please refer to TDengine Node.js connector for details.

Create connection

When using the Node.js connector, you must execute require td2.0-connector, and then use the taos.connect function. The parameter that taos.connect function must provide is host, and other parameters will use the following default values if they are not provided. Finally, the cursor needs to be initialized to communicate with the TDengine server-side.

const taos = require('td2.0-connector');
var conn = taos.connect({host:"", user:"root", password:"taosdata", config:"/etc/taos",port:0})
var cursor = conn.cursor(); // Initializing a new cursor

To close the connect:


To execute SQL and insert data

For DDL statements (such as create database, create table, use, and so on), you can use the execute method of cursor. The code is as follows:

cursor.execute('create database if not exists test;')

The above code creates a database named test. For DDL statements, there is generally no return value, and the execute return value of cursor is 0.

For Insert statements, the code is as follows:

var affectRows = cursor.execute('insert into values(now, 22.3, 34);')

The return value of the execute method is the number of rows affected by the statement. If the sql above inserts a piece of data into the weather table of the test database, the return value affectRows is 1.

TDengine does not currently support update and delete statements.


You can query the database through cursor.query function.

var query = cursor.query('show databases;')

The results of the query can be obtained and printed through query.execute() function:

var promise = query.execute();
promise.then(function(result) {

You can also use the bind method of query to format query statements. For example: query() automatically fills the ? with the value provided in the query statement .

var query = cursor.query('select * from meterinfo.meters where ts <= ? and areaid = ?;').bind(new Date(), 5);
query.execute().then(function(result) {

If you provide the second parameter in the query() statement and set it to true, you can also get the query results immediately. As follows:

var promise = cursor.query('select * from meterinfo.meters where v1 = 30;', true)
promise.then(function(result) {

Asynchronous function

The operation of asynchronous query database is similar to the above, only by adding _a after cursor.execute(), TaosQuery.execute() and other functions.

var promise1 = cursor.query('select count(*), avg(v1), avg(v2) from meter1;').execute_a()
var promise2 = cursor.query('select count(*), avg(v1), avg(v2) from meter2;').execute_a();
promise1.then(function(result) {
promise2.then(function(result) {


node-example.js provides a code example that uses the Node.js connector to create a table, insert weather data, and query the inserted data.

node-example-raw.js is also a code example that uses the Node.js connector to create a table, insert weather data, and query the inserted data, but unlike the above, this example only uses cursor.