使用Apache IoTDB进行IoT相关开发的架构设计与功能实现(12)

现在到了使用Apache IoTDB进行IoT相关开发的架构设计与功能实现的最后一个环境，在本文中我将向大家介绍IoTDB的查询语言。IoTDB为咱们广大开发者提供了类似SQL的查询语言，用于与IoTDB进行交互，查询语言可以分为4个主要部分：

• 架构语句：本节中列出了有关架构管理的语句。
• 数据管理语句：本节中列出了有关数据管理的语句（如：数据插入、数据查询等）。
• 数据库管理语句：本节中列出了有关数据库管理和身份验证的语句。
• 功能：IoTDB提供的功能均在本节中列出。

所有这些语句都是用IoTDB自己的语法编写的，有关语法组成的详细信息，请查看该部分。Reference

关键字

请不要使用这些关键字作为标识符。

CREATE, INSERT, UPDATE, DELETE, SELECT, SHOW, GRANT, INTO, SET, WHERE, FROM, TO, BY, DEVICE,
CONFIGURATION, DESCRIBE, SLIMIT, LIMIT, UNLINK, OFFSET, SOFFSET, FILL, LINEAR, PREVIOUS, PREVIOUSUNTILLAST,
METADATA, TIMESERIES, TIMESTAMP, PROPERTY, WITH, ROOT, DATATYPE, COMPRESSOR, STORAGE, GROUP, LABEL,INT32,
INT64, FLOAT, DOUBLE, BOOLEAN, TEXT, ENCODING, PLAIN, PLAIN_DICTIONARY, RLE, DIFF, TS_2DIFF, GORILLA, REGULAR,
BITMAP, ADD, UPSERT, VALUES, NOW, LINK, INDEX, USING, ON, DROP, MERGE, LIST, USER, PRIVILEGES, ROLE, ALL, OF,
ALTER, PASSWORD, REVOKE, LOAD, WATERMARK_EMBEDDING, UNSET, TTL, FLUSH, TASK, INFO, DYNAMIC, PARAMETER, VERSION,
REMOVE, MOVE, CHILD, PATHS, DEVICES, COUNT, NODES, LEVEL, MIN_TIME, MAX_TIME, MIN_VALUE, MAX_VALUE, AVG, FIRST_VALUE,
SUM, LAST_VALUE, LAST, DISABLE, ALIGN, COMPRESSION, TIME, ATTRIBUTES, TAGS,RENAME, FULL, CLEAR, CACHE

显示版本

show version

+---------------+
|        version|
+---------------+
|0.10.0-SNAPSHOT|
+---------------+
Total line number = 1
It costs 0.417s

架构语句

• 设置存储组

SET STORAGE GROUP TO <FullPath>
Eg: IoTDB > SET STORAGE GROUP TO root.ln.wf01.wt01
Note: FullPath can not include `*`

• 删除存储组

DELETE STORAGE GROUP <FullPath> [COMMA <FullPath>]*
Eg: IoTDB > DELETE STORAGE GROUP root.ln.wf01.wt01
Eg: IoTDB > DELETE STORAGE GROUP root.ln.wf01.wt01, root.ln.wf01.wt02
Note: FullPath can not include `*`

• 创建时间序列语句

CREATE TIMESERIES <FullPath> WITH <AttributeClauses>
alias
    : LR_BRACKET ID RR_BRACKET
    ;
attributeClauses
    : DATATYPE OPERATOR_EQ dataType COMMA ENCODING OPERATOR_EQ encoding
    (COMMA (COMPRESSOR | COMPRESSION) OPERATOR_EQ compressor=propertyValue)?
    (COMMA property)*
    tagClause
    attributeClause
    ;
attributeClause
    : (ATTRIBUTES LR_BRACKET property (COMMA property)* RR_BRACKET)?
    ;
tagClause
    : (TAGS LR_BRACKET property (COMMA property)* RR_BRACKET)?
    ;
DataTypeValue: BOOLEAN | DOUBLE | FLOAT | INT32 | INT64 | TEXT
EncodingValue: GORILLA | PLAIN | RLE | TS_2DIFF | REGULAR
CompressorValue: UNCOMPRESSED | SNAPPY
Eg: CREATE TIMESERIES root.ln.wf01.wt01.status WITH DATATYPE=BOOLEAN, ENCODING=PLAIN
Eg: CREATE TIMESERIES root.ln.wf01.wt01.temperature WITH DATATYPE=FLOAT, ENCODING=RLE
Eg: CREATE TIMESERIES root.ln.wf01.wt01.temperature WITH DATATYPE=FLOAT, ENCODING=RLE, COMPRESSOR=SNAPPY, MAX_POINT_NUMBER=3
Eg: create timeseries root.turbine.d0.s0(temperature) with datatype=FLOAT, encoding=RLE, compression=SNAPPY tags(unit=f, description='turbine this is a test1') attributes(H_Alarm=100, M_Alarm=50)
Note: Datatype and encoding type must be corresponding. Please check Chapter 3 Encoding Section for details.

• 删除时间序列语句

DELETE TIMESERIES <PrefixPath> [COMMA <PrefixPath>]*
Eg: IoTDB > DELETE TIMESERIES root.ln.wf01.wt01.status
Eg: IoTDB > DELETE TIMESERIES root.ln.wf01.wt01.status, root.ln.wf01.wt01.temperature
Eg: IoTDB > DELETE TIMESERIES root.ln.wf01.wt01.*

• 更改时间序列语句

ALTER TIMESERIES fullPath alterClause
alterClause
    : RENAME beforeName=ID TO currentName=ID
    | SET property (COMMA property)*
    | DROP ID (COMMA ID)*
    | ADD TAGS property (COMMA property)*
    | ADD ATTRIBUTES property (COMMA property)*
    | UPSERT tagClause attributeClause
    ;
attributeClause
    : (ATTRIBUTES LR_BRACKET property (COMMA property)* RR_BRACKET)?
    ;
tagClause
    : (TAGS LR_BRACKET property (COMMA property)* RR_BRACKET)?
    ;
Eg: ALTER timeseries root.turbine.d1.s1 RENAME tag1 TO newTag1
Eg: ALTER timeseries root.turbine.d1.s1 SET tag1=newV1, attr1=newV1
Eg: ALTER timeseries root.turbine.d1.s1 DROP tag1, tag2
Eg: ALTER timeseries root.turbine.d1.s1 ADD TAGS tag3=v3, tag4=v4
Eg: ALTER timeseries root.turbine.d1.s1 ADD ATTRIBUTES attr3=v3, attr4=v4
EG: ALTER timeseries root.turbine.d1.s1 UPSERT TAGS(tag2=newV2, tag3=v3) ATTRIBUTES(attr3=v3, attr4=v4)

• 显示所有时间序列语句

SHOW TIMESERIES
Eg: IoTDB > SHOW TIMESERIES
Note: This statement can only be used in IoTDB Client. If you need to show all timeseries in JDBC, please use `DataBaseMetadata` interface.

• 显示特定时间序列语句

SHOW TIMESERIES <Path>
Eg: IoTDB > SHOW TIMESERIES root
Eg: IoTDB > SHOW TIMESERIES root.ln
Eg: IoTDB > SHOW TIMESERIES root.ln.*.*.status
Eg: IoTDB > SHOW TIMESERIES root.ln.wf01.wt01.status
Note: The path can be prefix path, star path or timeseries path
Note: This statement can be used in IoTDB Client and JDBC.

• 显示带有 where 子句的特定时间序列语句

SHOW TIMESERIES prefixPath? showWhereClause?
showWhereClause
    : WHERE (property | containsExpression)
    ;
containsExpression
    : name=ID OPERATOR_CONTAINS value=propertyValue
    ;

Eg: show timeseries root.ln where unit='c'
Eg: show timeseries root.ln where description contains 'test1'

• 显示特定时间序列语句，其中子句从偏移量开始并限制结果总数

SHOW TIMESERIES prefixPath? showWhereClause? limitClause?

showWhereClause
    : WHERE (property | containsExpression)
    ;
containsExpression
    : name=ID OPERATOR_CONTAINS value=propertyValue
    ;
limitClause
    : LIMIT INT offsetClause?
    | offsetClause? LIMIT INT
    ;

Eg: show timeseries root.ln where unit='c'
Eg: show timeseries root.ln where description contains 'test1'
Eg: show timeseries root.ln where unit='c' limit 10 offset 10

• 显示存储组声明

SHOW STORAGE GROUP
Eg: IoTDB > SHOW STORAGE GROUP
Note: This statement can be used in IoTDB Client and JDBC.

• 计数时间序列语句

COUNT TIMESERIES <Path>
Eg: IoTDB > COUNT TIMESERIES root
Eg: IoTDB > COUNT TIMESERIES root.ln
Eg: IoTDB > COUNT TIMESERIES root.ln.*.*.status
Eg: IoTDB > COUNT TIMESERIES root.ln.wf01.wt01.status
Note: The path can be prefix path, star path or timeseries path.
Note: This statement can be used in IoTDB Client and JDBC.

COUNT TIMESERIES <Path> GROUP BY LEVEL=<INTEGER>
Eg: IoTDB > COUNT TIMESERIES root GROUP BY LEVEL=1
Eg: IoTDB > COUNT TIMESERIES root.ln GROUP BY LEVEL=2
Eg: IoTDB > COUNT TIMESERIES root.ln.wf01 GROUP BY LEVEL=3
Note: The path can be prefix path or timeseries path.
Note: This statement can be used in IoTDB Client and JDBC.

• 计数节点声明

COUNT NODES <Path> LEVEL=<INTEGER>
Eg: IoTDB > COUNT NODES root LEVEL=2
Eg: IoTDB > COUNT NODES root.ln LEVEL=2
Eg: IoTDB > COUNT NODES root.ln.wf01 LEVEL=3
Note: The path can be prefix path or timeseries path.
Note: This statement can be used in IoTDB Client and JDBC.

• 显示所有设备声明

SHOW Devices
Eg: IoTDB > SHOW Devices
Note: This statement can be used in IoTDB Client and JDBC.

• 显示特定设备声明

SHOW DEVICES <PrefixPath>
Eg: IoTDB > SHOW DEVICES root
Eg: IoTDB > SHOW DEVICES root.ln
Eg: IoTDB > SHOW DEVICES root.*.wf01
Note: The path can be prefix path or star path.
Note: This statement can be used in IoTDB Client and JDBC.

• 显示根语句的子路径

SHOW CHILD PATHS
Eg: IoTDB > SHOW CHILD PATHS
Note: This statement can be used in IoTDB Client and JDBC.

• 显示子路径声明

SHOW CHILD PATHS <Path>
Eg: IoTDB > SHOW CHILD PATHS root
Eg: IoTDB > SHOW CHILD PATHS root.ln
Eg: IoTDB > SHOW CHILD PATHS root.*.wf01
Eg: IoTDB > SHOW CHILD PATHS root.ln.wf*
Note: The path can be prefix path or star path, the nodes can be in a "prefix + star" format. 
Note: This statement can be used in IoTDB Client and JDBC.

数据管理声明

• 插入记录语句

INSERT INTO <PrefixPath> LPAREN TIMESTAMP COMMA <Sensor> [COMMA <Sensor>]* RPAREN VALUES LPAREN <TimeValue>, <PointValue> [COMMA <PointValue>]* RPAREN
Sensor : Identifier
Eg: IoTDB > INSERT INTO root.ln.wf01.wt01(timestamp,status) values(1509465600000,true)
Eg: IoTDB > INSERT INTO root.ln.wf01.wt01(timestamp,status) VALUES(NOW(), false)
Eg: IoTDB > INSERT INTO root.ln.wf01.wt01(timestamp,temperature) VALUES(2017-11-01T00:17:00.000+08:00,24.22028)
Eg: IoTDB > INSERT INTO root.ln.wf01.wt01(timestamp, status, temperature) VALUES (1509466680000, false, 20.060787);
Note: the statement needs to satisfy this constraint: <PrefixPath> + <Path> = <Timeseries>
Note: The order of Sensor and PointValue need one-to-one correspondence

• 更新记录声明

UPDATE <UpdateClause> SET <SetClause> WHERE <WhereClause>
UpdateClause: <prefixPath>
SetClause: <SetExpression> 
SetExpression: <Path> EQUAL <PointValue>
WhereClause : <Condition> [(AND | OR) <Condition>]*
Condition  : <Expression> [(AND | OR) <Expression>]*
Expression : [NOT | !]? TIME PrecedenceEqualOperator <TimeValue>
Eg: IoTDB > UPDATE root.ln.wf01.wt01 SET temperature = 23 WHERE time < NOW() and time > 2017-11-1T00:15:00+08:00
Note: the statement needs to satisfy this constraint: <PrefixPath> + <Path> = <Timeseries>

• 删除记录语句

DELETE FROM <PrefixPath> [COMMA <PrefixPath>]* WHERE TIME LESSTHAN <TimeValue>
Eg: DELETE FROM root.ln.wf01.wt01.temperature WHERE time < 2017-11-1T00:05:00+08:00
Eg: DELETE FROM root.ln.wf01.wt01.status, root.ln.wf01.wt01.temperature WHERE time < NOW()
Eg: DELETE FROM root.ln.wf01.wt01.* WHERE time < 1509466140000

• 选择记录语句

SELECT <SelectClause> FROM <FromClause> [WHERE <WhereClause>]?
SelectClause : <SelectPath> (COMMA <SelectPath>)*
SelectPath : <FUNCTION> LPAREN <Path> RPAREN | <Path>
FUNCTION : ‘COUNT’ , ‘MIN_TIME’, ‘MAX_TIME’, ‘MIN_VALUE’, ‘MAX_VALUE’
FromClause : <PrefixPath> (COMMA <PrefixPath>)?
WhereClause : <Condition> [(AND | OR) <Condition>]*
Condition  : <Expression> [(AND | OR) <Expression>]*
Expression : [NOT | !]? <TimeExpr> | [NOT | !]? <SensorExpr>
TimeExpr : TIME PrecedenceEqualOperator (<TimeValue> | <RelativeTime>)
RelativeTimeDurationUnit = Integer ('Y'|'MO'|'W'|'D'|'H'|'M'|'S'|'MS'|'US'|'NS')
RelativeTime : (now() | <TimeValue>) [(+|-) RelativeTimeDurationUnit]+
SensorExpr : (<Timeseries> | <Path>) PrecedenceEqualOperator <PointValue>
Eg: IoTDB > SELECT status, temperature FROM root.ln.wf01.wt01 WHERE temperature < 24 and time > 2017-11-1 0:13:00
Eg. IoTDB > SELECT * FROM root
Eg. IoTDB > SELECT * FROM root where time > now() - 5m
Eg. IoTDB > SELECT * FROM root.ln.*.wf*
Eg. IoTDB > SELECT COUNT(temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature < 25
Eg. IoTDB > SELECT MIN_TIME(temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature < 25
Eg. IoTDB > SELECT MAX_TIME(temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature > 24
Eg. IoTDB > SELECT MIN_VALUE(temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature > 23
Eg. IoTDB > SELECT MAX_VALUE(temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature < 25
Note: the statement needs to satisfy this constraint: <Path>(SelectClause) + <PrefixPath>(FromClause) = <Timeseries>
Note: If the <SensorExpr>(WhereClause) is started with <Path> and not with ROOT, the statement needs to satisfy this constraint: <PrefixPath>(FromClause) + <Path>(SensorExpr) = <Timeseries>
Note: In Version 0.7.0, if <WhereClause> includes `OR`, time filter can not be used.
Note: There must be a space on both sides of the plus and minus operator appearing in the time expression

• 分组依据语句

SELECT <SelectClause> FROM <FromClause> WHERE  <WhereClause> GROUP BY <GroupByClause>
SelectClause : <Function> [COMMA < Function >]*
Function : <AggregationFunction> LPAREN <Path> RPAREN
FromClause : <PrefixPath>
WhereClause : <Condition> [(AND | OR) <Condition>]*
Condition  : <Expression> [(AND | OR) <Expression>]*
Expression : [NOT | !]? <TimeExpr> | [NOT | !]? <SensorExpr>
TimeExpr : TIME PrecedenceEqualOperator (<TimeValue> | <RelativeTime>)
RelativeTimeDurationUnit = Integer ('Y'|'MO'|'W'|'D'|'H'|'M'|'S'|'MS'|'US'|'NS')
RelativeTime : (now() | <TimeValue>) [(+|-) RelativeTimeDurationUnit]+
SensorExpr : (<Timeseries> | <Path>) PrecedenceEqualOperator <PointValue>
GroupByClause : LPAREN <TimeInterval> COMMA <TimeUnit> (COMMA <TimeUnit>)? RPAREN
TimeInterval: LSBRACKET <TimeValue> COMMA <TimeValue> RRBRACKET | LRBRACKET <TimeValue> COMMA <TimeValue> RSBRACKET
TimeUnit : Integer <DurationUnit>
DurationUnit : "ms" | "s" | "m" | "h" | "d" | "w"
Eg: SELECT COUNT(status), COUNT(temperature) FROM root.ln.wf01.wt01 where temperature < 24 GROUP BY([1509465720000, 1509466380000), 5m)
Eg: SELECT COUNT(status), COUNT(temperature) FROM root.ln.wf01.wt01 where temperature < 24 GROUP BY((1509465720000, 1509466380000], 5m)
Eg. SELECT COUNT (status), MAX_VALUE(temperature) FROM root.ln.wf01.wt01 WHERE time < 1509466500000 GROUP BY([1509465720000, 1509466380000), 5m, 10m)
Eg. SELECT MIN_TIME(status), MIN_VALUE(temperature) FROM root.ln.wf01.wt01 WHERE temperature < 25 GROUP BY ([1509466140000, 1509466380000), 3m, 5ms)
Eg. SELECT MIN_TIME(status), MIN_VALUE(temperature) FROM root.ln.wf01.wt01 WHERE temperature < 25 GROUP BY ((1509466140000, 1509466380000], 3m, 5ms)
Note: the statement needs to satisfy this constraint: <Path>(SelectClause) + <PrefixPath>(FromClause) = <Timeseries>
Note: If the <SensorExpr>(WhereClause) is started with <Path> and not with ROOT, the statement needs to satisfy this constraint: <PrefixPath>(FromClause) + <Path>(SensorExpr) = <Timeseries>
Note: <TimeValue>(TimeInterval) needs to be greater than 0
Note: First <TimeValue>(TimeInterval) in needs to be smaller than second <TimeValue>(TimeInterval)
Note: <TimeUnit> needs to be greater than 0
Note: Third <TimeUnit> if set shouldn't be smaller than second <TimeUnit>

• 填写声明

SELECT <SelectClause> FROM <FromClause> WHERE <WhereClause> FILL <FillClause>
SelectClause : <Path> [COMMA <Path>]*
FromClause : < PrefixPath > [COMMA < PrefixPath >]*
WhereClause : <WhereExpression>
WhereExpression : TIME EQUAL <TimeValue>
FillClause : LPAREN <TypeClause> [COMMA <TypeClause>]* RPAREN
TypeClause : <Int32Clause> | <Int64Clause> | <FloatClause> | <DoubleClause> | <BoolClause> | <TextClause>
Int32Clause: INT32 LBRACKET (<LinearClause> | <PreviousClause>)  RBRACKET
Int64Clause: INT64 LBRACKET (<LinearClause> | <PreviousClause>)  RBRACKET
FloatClause: FLOAT LBRACKET (<LinearClause> | <PreviousClause>)  RBRACKET
DoubleClause: DOUBLE LBRACKET (<LinearClause> | <PreviousClause>)  RBRACKET
BoolClause: BOOLEAN LBRACKET (<LinearClause> | <PreviousClause>)  RBRACKET
TextClause: TEXT LBRACKET (<LinearClause> | <PreviousClause>)  RBRACKET
PreviousClause : PREVIOUS [COMMA <ValidPreviousTime>]?
LinearClause : LINEAR [COMMA <ValidPreviousTime> COMMA <ValidBehindTime>]?
ValidPreviousTime, ValidBehindTime: <TimeUnit>
TimeUnit : Integer <DurationUnit>
DurationUnit : "ms" | "s" | "m" | "h" | "d" | "w"
Eg: SELECT temperature FROM root.ln.wf01.wt01 WHERE time = 2017-11-01T16:37:50.000 FILL(float[previous, 1m])
Eg: SELECT temperature,status FROM root.ln.wf01.wt01 WHERE time = 2017-11-01T16:37:50.000 FILL (float[linear, 1m, 1m], boolean[previous, 1m])
Eg: SELECT temperature,status,hardware FROM root.ln.wf01.wt01 WHERE time = 2017-11-01T16:37:50.000 FILL (float[linear, 1m, 1m], boolean[previous, 1m], text[previous])
Eg: SELECT temperature,status,hardware FROM root.ln.wf01.wt01 WHERE time = 2017-11-01T16:37:50.000 FILL (float[linear], boolean[previous, 1m], text[previous])
Note: the statement needs to satisfy this constraint: <PrefixPath>(FromClause) + <Path>(SelectClause) = <Timeseries>
Note: Integer in <TimeUnit> needs to be greater than 0

• 分组依据填充语句

SELECT <SelectClause> FROM <FromClause> WHERE  <WhereClause> GROUP BY <GroupByClause> (FILL <GROUPBYFillClause>)?
GroupByClause : LPAREN <TimeInterval> COMMA <TimeUnit> RPAREN
GROUPBYFillClause : LPAREN <TypeClause> RPAREN
TypeClause : <AllClause> | <Int32Clause> | <Int64Clause> | <FloatClause> | <DoubleClause> | <BoolClause> | <TextClause> 
AllClause: ALL LBRACKET (<PreviousUntilLastClause> | <PreviousClause>)  RBRACKET
Int32Clause: INT32 LBRACKET (<PreviousUntilLastClause> | <PreviousClause>)  RBRACKET
Int64Clause: INT64 LBRACKET (<PreviousUntilLastClause> | <PreviousClause>)  RBRACKET
FloatClause: FLOAT LBRACKET (<PreviousUntilLastClause> | <PreviousClause>)  RBRACKET
DoubleClause: DOUBLE LBRACKET (<PreviousUntilLastClause> | <PreviousClause>)  RBRACKET
BoolClause: BOOLEAN LBRACKET (<PreviousUntilLastClause> | <PreviousClause>)  RBRACKET
TextClause: TEXT LBRACKET (<PreviousUntilLastClause> | <PreviousClause>)  RBRACKET
PreviousClause : PREVIOUS
PreviousUntilLastClause : PREVIOUSUNTILLAST
Eg: SELECT last_value(temperature) FROM root.ln.wf01.wt01 GROUP BY([20, 100), 5m) FILL (float[PREVIOUS])
Eg: SELECT last_value(temperature) FROM root.ln.wf01.wt01 GROUP BY((15, 100], 5m) FILL (float[PREVIOUS])
Eg: SELECT last_value(power) FROM root.ln.wf01.wt01 GROUP BY([20, 100), 5m) FILL (int32[PREVIOUSUNTILLAST])
Eg: SELECT last_value(power) FROM root.ln.wf01.wt01 GROUP BY([20, 100), 5m) FILL (int32[PREVIOUSUNTILLAST, 5m])
Eg: SELECT last_value(temperature), last_value(power) FROM root.ln.wf01.wt01 GROUP BY([20, 100), 5m) FILL (ALL[PREVIOUS])
Eg: SELECT last_value(temperature), last_value(power) FROM root.ln.wf01.wt01 GROUP BY([20, 100), 5m) FILL (ALL[PREVIOUS, 5m])
Note: In group by fill, sliding step is not supported in group by clause
Note: Now, only last_value aggregation function is supported in group by fill.
Note: Linear fill is not supported in group by fill.

• 限价声明

SELECT <SelectClause> FROM <FromClause> [WHERE <WhereClause>] [<LIMITClause>] [<SLIMITClause>]
SelectClause : [<Path> | Function]+
Function : <AggregationFunction> LPAREN <Path> RPAREN
FromClause : <Path>
WhereClause : <Condition> [(AND | OR) <Condition>]*
Condition : <Expression> [(AND | OR) <Expression>]*
Expression: [NOT|!]?<TimeExpr> | [NOT|!]?<SensorExpr>
TimeExpr : TIME PrecedenceEqualOperator (<TimeValue> | <RelativeTime>)
RelativeTimeDurationUnit = Integer ('Y'|'MO'|'W'|'D'|'H'|'M'|'S'|'MS'|'US'|'NS')
RelativeTime : (now() | <TimeValue>) [(+|-) RelativeTimeDurationUnit]+
SensorExpr : (<Timeseries>|<Path>) PrecedenceEqualOperator <PointValue>
LIMITClause : LIMIT <N> [OFFSETClause]?
N : Integer
OFFSETClause : OFFSET <OFFSETValue>
OFFSETValue : Integer
SLIMITClause : SLIMIT <SN> [SOFFSETClause]?
SN : Integer
SOFFSETClause : SOFFSET <SOFFSETValue>
SOFFSETValue : Integer
Eg: IoTDB > SELECT status, temperature FROM root.ln.wf01.wt01 WHERE temperature < 24 and time > 2017-11-1 0:13:00 LIMIT 3 OFFSET 2
Eg. IoTDB > SELECT COUNT (status), MAX_VALUE(temperature) FROM root.ln.wf01.wt01 WHERE time < 1509466500000 GROUP BY([1509465720000, 1509466380000], 5m) LIMIT 3
Note: N, OFFSETValue, SN and SOFFSETValue must be greater than 0.
Note: The order of <LIMITClause> and <SLIMITClause> does not affect the grammatical correctness.
Note: <FillClause> can not use <LIMITClause> but not <SLIMITClause>.

• 按设备对齐语句

AlignbyDeviceClause : ALIGN BY DEVICE

Rules:  
1. Both uppercase and lowercase are ok.  
Correct example: select * from root.sg1 align by device  
Correct example: select * from root.sg1 ALIGN BY DEVICE  

2. AlignbyDeviceClause can only be used at the end of a query statement.  
Correct example: select * from root.sg1 where time > 10 align by device  
Wrong example: select * from root.sg1 align by device where time > 10  

3. The paths of the SELECT clause can only be single level. In other words, the paths of the SELECT clause can only be measurements or STAR, without DOT.
Correct example: select s0,s1 from root.sg1.* align by device  
Correct example: select s0,s1 from root.sg1.d0, root.sg1.d1 align by device  
Correct example: select * from root.sg1.* align by device  
Correct example: select * from root align by device  
Correct example: select s0,s1,* from root.*.* align by device  
Wrong example: select d0.s1, d0.s2, d1.s0 from root.sg1 align by device  
Wrong example: select *.s0, *.s1 from root.* align by device  
Wrong example: select *.*.* from root align by device

4. The data types of the same measurement column should be the same across devices. 
Note that when it comes to aggregated paths, the data type of the measurement column will reflect 
the aggregation function rather than the original timeseries.

Correct example: select s0 from root.sg1.d0,root.sg1.d1 align by device   
root.sg1.d0.s0 and root.sg1.d1.s0 are both INT32.  

Correct example: select count(s0) from root.sg1.d0,root.sg1.d1 align by device   
count(root.sg1.d0.s0) and count(root.sg1.d1.s0) are both INT64.  

Wrong example: select s0 from root.sg1.d0, root.sg2.d3 align by device  
root.sg1.d0.s0 is INT32 while root.sg2.d3.s0 is FLOAT. 

5. The display principle of the result table is that all the columns (no matther whther a column has has existing data) will be shown, with nonexistent cells being null. Besides, the select clause support const column (e.g., 'a', '123' etc..).  
For example, "select s0,s1,s2,'abc',s1,s2 from root.sg.d0, root.sg.d1, root.sg.d2 align by device". Suppose that the actual existing timeseries are as follows:  
- root.sg.d0.s0
- root.sg.d0.s1
- root.sg.d1.s0

Then you could expect a table like:  

| Time | Device   | s0 | s1 |  s2  | 'abc' | s1 |  s2  |
| ---  | ---      | ---| ---| null | 'abc' | ---| null |
|  1   |root.sg.d0| 20 | 2.5| null | 'abc' | 2.5| null |
|  2   |root.sg.d0| 23 | 3.1| null | 'abc' | 3.1| null |
| ...  | ...      | ...| ...| null | 'abc' | ...| null |
|  1   |root.sg.d1| 12 |null| null | 'abc' |null| null |
|  2   |root.sg.d1| 19 |null| null | 'abc' |null| null |
| ...  | ...      | ...| ...| null | 'abc' | ...| null |

Note that the cells of measurement 's0' and device 'root.sg.d1' are all null.    

6. The duplicated devices in the prefix paths are neglected.  
For example, "select s0,s1 from root.sg.d0,root.sg.d0,root.sg.d1 align by device" is equal to "select s0,s1 from root.sg.d0,root.sg.d1 align by device".  
For example. "select s0,s1 from root.sg.*,root.sg.d0 align by device" is equal to "select s0,s1 from root.sg.* align by device".  

7. The duplicated measurements in the suffix paths are not neglected.  
For example, "select s0,s0,s1 from root.sg.* align by device" is not equal to "select s0,s1 from root.sg.* align by device".

8. Both time predicates and value predicates are allowed in Where Clause. The paths of the value predicates can be the leaf node or full path started with ROOT. And wildcard is not allowed here. For example:
- select * from root.sg.* where time = 1 align by device
- select * from root.sg.* where s0 < 100 align by device
- select * from root.sg.* where time < 20 AND s0 > 50 align by device
- select * from root.sg.d0 where root.sg.d0.s0 = 15 align by device

9. More correct examples:
   - select * from root.vehicle align by device
   - select s0,s0,s1 from root.vehicle.* align by device
   - select s0,s1 from root.vehicle.* limit 10 offset 1 align by device
   - select * from root.vehicle slimit 10 soffset 2 align by device
   - select * from root.vehicle where time > 10 align by device
   - select * from root.vehicle.* where time < 10 AND s0 > 25 align by device
   - select * from root.vehicle where root.vehicle.d0.s0>0 align by device
   - select count(*) from root.vehicle align by device
   - select sum(*) from root.vehicle GROUP BY (20ms,0,[2,50]) align by device
   - select * from root.vehicle where time = 3 Fill(int32[previous, 5ms]) align by device

• 禁用对齐语句

Disable Align Clause: DISABLE ALIGN

Rules:  
1. Both uppercase and lowercase are ok.  
Correct example: select * from root.sg1 disable align  
Correct example: select * from root.sg1 DISABLE ALIGN  

2. Disable Align Clause can only be used at the end of a query statement.  
Correct example: select * from root.sg1 where time > 10 disable align 
Wrong example: select * from root.sg1 disable align where time > 10 

3. Disable Align Clause cannot be used with Aggregation, Fill Statements, Group By or Group By Device Statements, but can with Limit Statements.
Correct example: select * from root.sg1 limit 3 offset 2 disable align
Correct example: select * from root.sg1 slimit 3 soffset 2 disable align
Wrong example: select count(s0),count(s1) from root.sg1.d1 disable align
Wrong example: select * from root.vehicle where root.vehicle.d0.s0>0 disable align
Wrong example: select * from root.vehicle align by device disable align

4. The display principle of the result table is that only when the column (or row) has existing data will the column (or row) be shown, with nonexistent cells being empty.

You could expect a table like:
| Time | root.sg.d0.s1 | Time | root.sg.d0.s2 | Time | root.sg.d1.s1 |
| ---  | ---           | ---  | ---           | ---  | ---           |
|  1   | 100           | 20   | 300           | 400  | 600           |
|  2   | 300           | 40   | 800           | 700  | 900           |
|  4   | 500           |      |               | 800  | 1000          |
|      |               |      |               | 900  | 8000          |

5. More correct examples: 
   - select * from root.vehicle disable align
   - select s0,s0,s1 from root.vehicle.* disable align
   - select s0,s1 from root.vehicle.* limit 10 offset 1 disable align
   - select * from root.vehicle slimit 10 soffset 2 disable align
   - select * from root.vehicle where time > 10 disable align

• 选择最后记录语句

LAST 函数返回给定时间序列的最后一个时间值对。目前，LAST 查询不支持筛选器。

SELECT LAST <SelectClause> FROM <FromClause>
Select Clause : <Path> [COMMA <Path>]*
FromClause : < PrefixPath > [COMMA < PrefixPath >]*

Eg. SELECT LAST s1 FROM root.sg.d1
Eg. SELECT LAST s1, s2 FROM root.sg.d1
Eg. SELECT LAST s1 FROM root.sg.d1, root.sg.d2

Rules:
1. the statement needs to satisfy this constraint: <PrefixPath> + <Path> = <Timeseries>

2. The result set of last query will always be displayed in a fixed three column table format.
For example, "select last s1, s2 from root.sg.d1, root.sg.d2", the query result would be:

| Time | Path         | Value |
| ---  | ------------ | ----- |
|  5   | root.sg.d1.s1| 100   |
|  2   | root.sg.d1.s2| 400   |
|  4   | root.sg.d2.s1| 250   |
|  9   | root.sg.d2.s2| 600   |

3. It is not supported to use "diable align" in LAST query.

数据库管理声明

• 创建用户

CREATE USER <userName> <password>;  
userName:=identifier  
password:=string
Eg: IoTDB > CREATE USER thulab 'pwd';

• 删除用户

DROP USER <userName>;  
userName:=identifier
Eg: IoTDB > DROP USER xiaoming;

• 创建角色

CREATE ROLE <roleName>;  
roleName:=identifie
Eg: IoTDB > CREATE ROLE admin;

• 删除角色

DROP ROLE <roleName>;  
roleName:=identifier
Eg: IoTDB > DROP ROLE admin;

• 授予用户权限

GRANT USER <userName> PRIVILEGES <privileges> ON <nodeName>;  
userName:=identifier  
nodeName:=identifier (DOT identifier)*  
privileges:= string (COMMA string)*
Eg: IoTDB > GRANT USER tempuser PRIVILEGES 'DELETE_TIMESERIES' on root.ln;

• 授予角色权限

GRANT ROLE <roleName> PRIVILEGES <privileges> ON <nodeName>;  
privileges:= string (COMMA string)*  
roleName:=identifier  
nodeName:=identifier (DOT identifier)*
Eg: IoTDB > GRANT ROLE temprole PRIVILEGES 'DELETE_TIMESERIES' ON root.ln;

• 授予用户角色

GRANT <roleName> TO <userName>;  
roleName:=identifier  
userName:=identifier
Eg: IoTDB > GRANT temprole TO tempuser;

• 撤销用户权限

REVOKE USER <userName> PRIVILEGES <privileges> ON <nodeName>;   
privileges:= string (COMMA string)*  
userName:=identifier  
nodeName:=identifier (DOT identifier)*
Eg: IoTDB > REVOKE USER tempuser PRIVILEGES 'DELETE_TIMESERIES' on root.ln;

• 撤销角色特权

REVOKE ROLE <roleName> PRIVILEGES <privileges> ON <nodeName>;  
privileges:= string (COMMA string)*  
roleName:= identifier  
nodeName:=identifier (DOT identifier)*
Eg: IoTDB > REVOKE ROLE temprole PRIVILEGES 'DELETE_TIMESERIES' ON root.ln;

• 撤消用户的角色

REVOKE <roleName> FROM <userName>;
roleName:=identifier
userName:=identifier
Eg: IoTDB > REVOKE temproleFROM tempuser;

• 列出用户

LIST USER
Eg: IoTDB > LIST USER

• 列出角色

LIST ROLE
Eg: IoTDB > LIST ROLE

• 列出权限

LIST PRIVILEGES USER  <username> ON <path>;    
username:=identifier    
path=‘root’ (DOT identifier)*
Eg: IoTDB > LIST PRIVIEGES USER sgcc_wirte_user ON root.sgcc;

• 列出角色权限（在特定路径上）

LIST PRIVILEGES ROLE <roleName> ON <path>;    
roleName:=identifier  
path=‘root’ (DOT identifier)*
Eg: IoTDB > LIST PRIVIEGES ROLE wirte_role ON root.sgcc;

• 列出用户的权限

LIST USER PRIVILEGES <username> ;   
username:=identifier  
Eg: IoTDB > LIST USER PRIVIEGES tempuser;

• 列出角色的权限

LIST ROLE PRIVILEGES <roleName>
roleName:=identifier
Eg: IoTDB > LIST ROLE PRIVIEGES actor;

• 列出用户的角色

LIST ALL ROLE OF USER <username> ;  
username:=identifier
Eg: IoTDB > LIST ALL ROLE OF USER tempuser;

• 列出角色的用户

LIST ALL USER OF ROLE <roleName>;
roleName:=identifier
Eg: IoTDB > LIST ALL USER OF ROLE roleuser;

• 更改密码

ALTER USER <username> SET PASSWORD <password>;
roleName:=identifier
password:=identifier
Eg: IoTDB > ALTER USER tempuser SET PASSWORD 'newpwd';

功能

• 计数

COUNT 函数返回由 SELECT 语句选择的时间序列（一个或多个）非空值的值数。结果是一个有符号的 64 位整数。如果没有匹配的行，则 COUNT （） 返回 0。

SELECT COUNT(Path) (COMMA COUNT(Path))* FROM <FromClause> [WHERE <WhereClause>]?
Eg. SELECT COUNT(status), COUNT(temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature < 24
Note: the statement needs to satisfy this constraint: <PrefixPath> + <Path> = <Timeseries>

• FIRST_VALUE（从FIRSTV0.10.0)

FIRST_VALUE函数返回所选时间序列的第一个点值（一个或多个）。

SELECT FIRST_VALUE (Path) (COMMA FIRST_VALUE (Path))* FROM <FromClause> [WHERE <WhereClause>]?
Eg. SELECT FIRST_VALUE (status), FIRST_VALUE (temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature < 24
Note: the statement needs to satisfy this constraint: <PrefixPath> + <Path> = <Timeseries>

• LAST_VALUE

LAST_VALUE函数返回所选时间序列的最后一个点值（一个或多个）。

SELECT LAST_VALUE (Path) (COMMA LAST_VALUE (Path))* FROM <FromClause> [WHERE <WhereClause>]?
Eg. SELECT LAST_VALUE (status), LAST_VALUE (temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature < 24
Note: the statement needs to satisfy this constraint: <PrefixPath> + <Path> = <Timeseries>

• MAX_TIME

MAX_TIME函数返回所选时间序列的最大时间戳（一个或多个）。结果是一个有符号的 64 位整数，大于 0。

SELECT MAX_TIME (Path) (COMMA MAX_TIME (Path))* FROM <FromClause> [WHERE <WhereClause>]?
Eg. SELECT MAX_TIME(status), MAX_TIME(temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature < 24
Note: the statement needs to satisfy this constraint: <PrefixPath> + <Path> = <Timeseries>

• MAX_VALUE

MAX_VALUE 函数返回所选时间序列（一个或多个）的最大值（按字典顺序排列）。

SELECT MAX_VALUE (Path) (COMMA MAX_VALUE (Path))* FROM <FromClause> [WHERE <WhereClause>]?
Eg. SELECT MAX_VALUE(status), MAX_VALUE(temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature < 24
Note: the statement needs to satisfy this constraint: <PrefixPath> + <Path> = <Timeseries>

• AVG（从MEANV0.9.0)

AVG 函数返回指定时间段内所选时间序列的算术平均值。时间序列必须是 int32、int64、float、double 类型，其他类型不计算。结果是一个双精度类型数字。

SELECT AVG (Path) (COMMA AVG (Path))* FROM <FromClause> [WHERE <WhereClause>]?
Eg. SELECT AVG (temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature < 24
Note: the statement needs to satisfy this constraint: <PrefixPath> + <Path> = <Timeseries>

• MIN_TIME

MIN_TIME函数返回所选时间序列（一个或多个）的最小时间戳。结果是一个有符号的 64 位整数，大于 0。

SELECT MIN_TIME (Path) (COMMA MIN_TIME (Path))*FROM <FromClause> [WHERE <WhereClause>]?
Eg. SELECT MIN_TIME(status), MIN_TIME(temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature < 24
Note: the statement needs to satisfy this constraint: <PrefixPath> + <Path> = <Timeseries>

• MIN_VALUE

MIN_VALUE函数返回所选时间序列（一个或多个）的最小值（按字典顺序排列）。

SELECT MIN_VALUE (Path) (COMMA MIN_VALUE (Path))* FROM <FromClause> [WHERE <WhereClause>]?
Eg. SELECT MIN_VALUE(status),MIN_VALUE(temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature < 24
Note: the statement needs to satisfy this constraint: <PrefixPath> + <Path> = <Timeseries>

• 现在

NOW 函数返回当前时间戳。此函数可以在数据操作语句中用于表示时间。结果是一个有符号的 64 位整数，大于 0。

NOW()
Eg. INSERT INTO root.ln.wf01.wt01(timestamp,status) VALUES(NOW(), false) 
Eg. UPDATE root.ln.wf01.wt01 SET temperature = 23 WHERE time < NOW()
Eg. DELETE FROM root.ln.wf01.wt01.status, root.ln.wf01.wt01.temperature WHERE time < NOW()
Eg. SELECT * FROM root WHERE time < NOW()
Eg. SELECT COUNT(temperature) FROM root.ln.wf01.wt01 WHERE time < NOW()

• 和

SUM 函数返回指定时间段内所选时间序列（一个或多个）的总和。时间序列必须是 int32、int64、float、double 类型，其他类型不计算。结果是一个双精度类型数字。

SELECT SUM(Path) (COMMA SUM(Path))* FROM <FromClause> [WHERE <WhereClause>]?
Eg. SELECT SUM(temperature) FROM root.ln.wf01.wt01 WHERE root.ln.wf01.wt01.temperature < 24
Note: the statement needs to satisfy this constraint: <PrefixPath> + <Path> = <Timeseries>

TTL

IoTDB支持存储级TTL设置，这意味着它能够自动和定期删除旧数据。使用 TTL 的好处是，希望您可以控制总磁盘空间使用量并防止计算机耗尽磁盘。此外，查询性能可能会随着文件总数的增加而下降，并且随着文件数量的增加，内存使用量也会增加。及时删除此类文件有助于保持较高的查询性能级别并减少内存使用量。IoTDB 中的 TTL 操作由以下三个语句支持：

• 设置 TTL

SET TTL TO StorageGroupName TTLTime
Eg. SET TTL TO root.group1 3600000
This example means that for data in root.group1, only that of the latest 1 hour will remain, the
older one is removed or made invisible. 
Note: TTLTime should be millisecond timestamp. When TTL is set, insertions that fall
out of TTL will be rejected.

• 未设置 TTL

UNSET TTL TO StorageGroupName
Eg. UNSET TTL TO root.group1
This example means that data of all time will be accepted in this group.

• 显示 TTL

SHOW ALL TTL
SHOW TTL ON StorageGroupNames
Eg.1 SHOW ALL TTL
This example will show TTLs of all storage groups.
Eg.2 SHOW TTL ON root.group1,root.group2,root.group3
This example will show TTLs of the specified 3 groups.
Notice: storage groups without TTL will show a "null"

注意：将 TTL 设置为某些存储组时，TTL 外的数据将立即变得不可见，但由于数据文件可能包含过时数据和活动数据，或者数据文件可能正被查询使用，因此数据的物理删除是过时的。如果在之前设置TTL后增加或取消设置TTL，则可能会再次看到一些以前不可见的数据，但物理删除的数据将永远丢失。换句话说，与删除语句不同，出于效率考虑，不保证数据删除的原子性。因此，建议您不要在设置TTL后更改它，或者至少不要经常重置它，除非您决心遭受不可预测性。

参考

关键字

Keywords for IoTDB (case insensitive):
ADD, BY, COMPRESSOR, CREATE, DATATYPE, DELETE, DESCRIBE, DROP, ENCODING, EXIT, FROM, GRANT, GROUP, LABLE, LINK, INDEX, INSERT, INTO, LOAD, MAX_POINT_NUMBER, MERGE, METADATA, ON, ORDER, PASSWORD, PRIVILEGES, PROPERTY, QUIT, REVOKE, ROLE, ROOT, SELECT, SET, SHOW, STORAGE, TIME, TIMESERIES, TIMESTAMP, TO, UNLINK, UPDATE, USER, USING, VALUE, VALUES, WHERE, WITH

Keywords with special meanings (case insensitive):
* Data Types: BOOLEAN, DOUBLE, FLOAT, INT32, INT64, TEXT 
* Encoding Methods: BITMAP, DFT, GORILLA, PLAIN, RLE, TS_2DIFF 
* Compression Methods: UNCOMPRESSED, SNAPPY, GZIP, LZ0, ZDT, PAA, PLA
* Logical symbol: AND, &, &&, OR, | , ||, NOT, !, TRUE, FALSE

标识符

QUOTE := '\'';
DOT := '.';
COLON : ':' ;
COMMA := ',' ;
SEMICOLON := ';' ;
LPAREN := '(' ;
RPAREN := ')' ;
LBRACKET := '[';
RBRACKET := ']';
EQUAL := '=' | '==';
NOTEQUAL := '<>' | '!=';
LESSTHANOREQUALTO := '<=';
LESSTHAN := '<';
GREATERTHANOREQUALTO := '>=';
GREATERTHAN := '>';
DIVIDE := '/';
PLUS := '+';
MINUS := '-';
STAR := '*';
Letter := 'a'..'z' | 'A'..'Z';
HexDigit := 'a'..'f' | 'A'..'F';
Digit := '0'..'9';
Boolean := TRUE | FALSE | 0 | 1 (case insensitive)

StringLiteral := ( '\'' ( ~('\'') )* '\'' | '\"' ( ~('\"') )* '\"');
eg. 'abc'
eg. 'abc'

Integer := ('-' | '+')? Digit+;
eg. 123
eg. -222

Float := ('-' | '+')? Digit+ DOT Digit+ (('e' | 'E') ('-' | '+')? Digit+)?;
eg. 3.1415
eg. 1.2E10
eg. -1.33

Identifier := (Letter | '_') (Letter | Digit | '_' | MINUS)*;
eg. a123
eg. _abc123

文字

PointValue : Integer | Float | StringLiteral | Boolean

TimeValue : Integer | DateTime | ISO8601 | NOW()
Note: Integer means timestamp type.

DateTime : 
eg. 2016-11-16T16:22:33+08:00
eg. 2016-11-16 16:22:33+08:00
eg. 2016-11-16T16:22:33.000+08:00
eg. 2016-11-16 16:22:33.000+08:00
Note: DateTime Type can support several types, see Chapter 3 Datetime section for details.

PrecedenceEqualOperator : EQUAL | NOTEQUAL | LESSTHANOREQUALTO | LESSTHAN | GREATERTHANOREQUALTO | GREATERTHAN

Timeseries : ROOT [DOT <LayerName>]* DOT <SensorName>
LayerName : Identifier
SensorName : Identifier
eg. root.ln.wf01.wt01.status
eg. root.sgcc.wf03.wt01.temperature
Note: Timeseries must be start with `root`(case insensitive) and end with sensor name.

PrefixPath : ROOT (DOT <LayerName>)*
LayerName : Identifier | STAR
eg. root.sgcc
eg. root.*

Path: (ROOT | <LayerName>) (DOT <LayerName>)* 
LayerName: Identifier | STAR
eg. root.ln.wf01.wt01.status
eg. root.*.wf01.wt01.status
eg. root.ln.wf01.wt01.*
eg. *.wt01.*
eg. *