docs/UserGuide/Operate-Metadata/Timeseries.md

<!--

    Licensed to the Apache Software Foundation (ASF) under one
    or more contributor license agreements.  See the NOTICE file
    distributed with this work for additional information
    regarding copyright ownership.  The ASF licenses this file
    to you under the Apache License, Version 2.0 (the
    "License"); you may not use this file except in compliance
    with the License.  You may obtain a copy of the License at
    
        http://www.apache.org/licenses/LICENSE-2.0
    
    Unless required by applicable law or agreed to in writing,
    software distributed under the License is distributed on an
    "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
    KIND, either express or implied.  See the License for the
    specific language governing permissions and limitations
    under the License.

-->

# Timeseries Management

## Create Timeseries

According to the storage model selected before, we can create corresponding timeseries in the two databases respectively. The SQL statements for creating timeseries are as follows:

```
IoTDB > create timeseries root.ln.wf01.wt01.status with datatype=BOOLEAN,encoding=PLAIN
IoTDB > create timeseries root.ln.wf01.wt01.temperature with datatype=FLOAT,encoding=RLE
IoTDB > create timeseries root.ln.wf02.wt02.hardware with datatype=TEXT,encoding=PLAIN
IoTDB > create timeseries root.ln.wf02.wt02.status with datatype=BOOLEAN,encoding=PLAIN
IoTDB > create timeseries root.sgcc.wf03.wt01.status with datatype=BOOLEAN,encoding=PLAIN
IoTDB > create timeseries root.sgcc.wf03.wt01.temperature with datatype=FLOAT,encoding=RLE
```

From v0.13, you can use a simplified version of the SQL statements to create timeseries:

```
IoTDB > create timeseries root.ln.wf01.wt01.status BOOLEAN encoding=PLAIN
IoTDB > create timeseries root.ln.wf01.wt01.temperature FLOAT encoding=RLE
IoTDB > create timeseries root.ln.wf02.wt02.hardware TEXT encoding=PLAIN
IoTDB > create timeseries root.ln.wf02.wt02.status BOOLEAN encoding=PLAIN
IoTDB > create timeseries root.sgcc.wf03.wt01.status BOOLEAN encoding=PLAIN
IoTDB > create timeseries root.sgcc.wf03.wt01.temperature FLOAT encoding=RLE
```

Notice that when in the CREATE TIMESERIES statement the encoding method conflicts with the data type, the system gives the corresponding error prompt as shown below:

```
IoTDB > create timeseries root.ln.wf02.wt02.status WITH DATATYPE=BOOLEAN, ENCODING=TS_2DIFF
error: encoding TS_2DIFF does not support BOOLEAN
```

Please refer to [Encoding](../Data-Concept/Encoding.md) for correspondence between data type and encoding.

## Create Aligned Timeseries

The SQL statement for creating a group of timeseries are as follows:

```
IoTDB> CREATE ALIGNED TIMESERIES root.ln.wf01.GPS(latitude FLOAT encoding=PLAIN compressor=SNAPPY, longitude FLOAT encoding=PLAIN compressor=SNAPPY)
```

You can set different datatype, encoding, and compression for the timeseries in a group of aligned timeseries

It is also supported to set an alias, tag, and attribute for aligned timeseries.

## Delete Timeseries

To delete the timeseries we created before, we are able to use `(DELETE | DROP) TimeSeries <PathPattern>` statement.

The usage are as follows:

```
IoTDB> delete timeseries root.ln.wf01.wt01.status
IoTDB> delete timeseries root.ln.wf01.wt01.temperature, root.ln.wf02.wt02.hardware
IoTDB> delete timeseries root.ln.wf02.*
IoTDB> drop timeseries root.ln.wf02.*
```

## Show Timeseries

* SHOW LATEST? TIMESERIES pathPattern? whereClause? limitClause?

  There are four optional clauses added in SHOW TIMESERIES, return information of time series 
  

Timeseries information includes: timeseries path, alias of measurement, database it belongs to, data type, encoding type, compression type, tags and attributes.

Examples:

* SHOW TIMESERIES

  presents all timeseries information in JSON form

* SHOW TIMESERIES <`PathPattern`> 
  
  returns all timeseries information matching the given <`PathPattern`>. SQL statements are as follows:

```
IoTDB> show timeseries root.**
IoTDB> show timeseries root.ln.**
```

The results are shown below respectively:

```
+-------------------------------+--------+-------------+--------+--------+-----------+-------------------------------------------+--------------------------------------------------------+--------+-------------------+
|                     timeseries|   alias|     database|dataType|encoding|compression|                                       tags|                                              attributes|deadband|deadband parameters|
+-------------------------------+--------+-------------+--------+--------+-----------+-------------------------------------------+--------------------------------------------------------+--------+-------------------+
|root.sgcc.wf03.wt01.temperature|    null|    root.sgcc|   FLOAT|     RLE|     SNAPPY|                                       null|                                                    null|    null|               null|
|     root.sgcc.wf03.wt01.status|    null|    root.sgcc| BOOLEAN|   PLAIN|     SNAPPY|                                       null|                                                    null|    null|               null|
|             root.turbine.d1.s1|newAlias| root.turbine|   FLOAT|     RLE|     SNAPPY|{"newTag1":"newV1","tag4":"v4","tag3":"v3"}|{"attr2":"v2","attr1":"newV1","attr4":"v4","attr3":"v3"}|    null|               null|
|     root.ln.wf02.wt02.hardware|    null|      root.ln|    TEXT|   PLAIN|     SNAPPY|                                       null|                                                    null|    null|               null|
|       root.ln.wf02.wt02.status|    null|      root.ln| BOOLEAN|   PLAIN|     SNAPPY|                                       null|                                                    null|    null|               null|
|  root.ln.wf01.wt01.temperature|    null|      root.ln|   FLOAT|     RLE|     SNAPPY|                                       null|                                                    null|    null|               null|
|       root.ln.wf01.wt01.status|    null|      root.ln| BOOLEAN|   PLAIN|     SNAPPY|                                       null|                                                    null|    null|               null|
+-------------------------------+--------+-------------+--------+--------+-----------+-------------------------------------------+--------------------------------------------------------+--------+-------------------+
Total line number = 7
It costs 0.016s

+-----------------------------+-----+-------------+--------+--------+-----------+----+----------+--------+-------------------+
|                   timeseries|alias|     database|dataType|encoding|compression|tags|attributes|deadband|deadband parameters|
+-----------------------------+-----+-------------+--------+--------+-----------+----+----------+--------+-------------------+
|   root.ln.wf02.wt02.hardware| null|      root.ln|    TEXT|   PLAIN|     SNAPPY|null|      null|    null|               null|
|     root.ln.wf02.wt02.status| null|      root.ln| BOOLEAN|   PLAIN|     SNAPPY|null|      null|    null|               null|
|root.ln.wf01.wt01.temperature| null|      root.ln|   FLOAT|     RLE|     SNAPPY|null|      null|    null|               null|
|     root.ln.wf01.wt01.status| null|      root.ln| BOOLEAN|   PLAIN|     SNAPPY|null|      null|    null|               null|
+-----------------------------+-----+-------------+--------+--------+-----------+----+----------+--------+-------------------+
Total line number = 4
It costs 0.004s
```

* SHOW TIMESERIES LIMIT INT OFFSET INT

  returns all the timeseries information start from the offset and limit the number of series returned. For example,

```
show timeseries root.ln.** limit 10 offset 10
```

* SHOW TIMESERIES WHERE TIMESERIES contains 'containStr'

  The query result set is filtered by string fuzzy matching based on the names of the timeseries. For example:

```
show timeseries root.ln.** where timeseries contains 'wf01.wt'
```

The result is shown below:

```
+-------------------------------+--------+-------------+--------+--------+-----------+-------------------------------------------+--------------------------------------------------------+--------+-------------------+
|                     timeseries|   alias|     database|dataType|encoding|compression|                                       tags|                                              attributes|deadband|deadband parameters|
+-------------------------------+--------+-------------+--------+--------+-----------+-------------------------------------------+--------------------------------------------------------+--------+-------------------+
|  root.ln.wf01.wt01.temperature|    null|      root.ln|   FLOAT|     RLE|     SNAPPY|                                       null|                                                    null|    null|               null|
|       root.ln.wf01.wt01.status|    null|      root.ln| BOOLEAN|   PLAIN|     SNAPPY|                                       null|                                                    null|    null|               null|
+-------------------------------+--------+-------------+--------+--------+-----------+-------------------------------------------+--------------------------------------------------------+--------+-------------------+
Total line number = 2
It costs 0.016s
```

* SHOW TIMESERIES WHERE DataType=type

  The query result set is filtered by data type. For example:

```
show timeseries root.ln.** where dataType=FLOAT
```

The result is shown below:

```
+-------------------------------+--------+-------------+--------+--------+-----------+-------------------------------------------+--------------------------------------------------------+--------+-------------------+
|                     timeseries|   alias|     database|dataType|encoding|compression|                                       tags|                                              attributes|deadband|deadband parameters|
+-------------------------------+--------+-------------+--------+--------+-----------+-------------------------------------------+--------------------------------------------------------+--------+-------------------+
|root.sgcc.wf03.wt01.temperature|    null|    root.sgcc|   FLOAT|     RLE|     SNAPPY|                                       null|                                                    null|    null|               null|
|             root.turbine.d1.s1|newAlias| root.turbine|   FLOAT|     RLE|     SNAPPY|{"newTag1":"newV1","tag4":"v4","tag3":"v3"}|{"attr2":"v2","attr1":"newV1","attr4":"v4","attr3":"v3"}|    null|               null|
|  root.ln.wf01.wt01.temperature|    null|      root.ln|   FLOAT|     RLE|     SNAPPY|                                       null|                                                    null|    null|               null|
+-------------------------------+--------+-------------+--------+--------+-----------+-------------------------------------------+--------------------------------------------------------+--------+-------------------+
Total line number = 3
It costs 0.016s

```


* SHOW LATEST TIMESERIES

  all the returned timeseries information should be sorted in descending order of the last timestamp of timeseries
  

It is worth noting that when the queried path does not exist, the system will return no timeseries.  


## Count Timeseries

IoTDB is able to use `COUNT TIMESERIES <Path>` to count the number of timeseries matching the path. SQL statements are as follows:
* `WHERE` condition could be used to fuzzy match a time series name with the following syntax: `COUNT TIMESERIES <Path> WHERE TIMESERIES contains 'containStr'`.
* `WHERE` condition could be used to filter result by data type with the syntax: `COUNT TIMESERIES <Path> WHERE DataType=<DataType>'`.
* `WHERE` condition could be used to filter result by tags with the syntax: `COUNT TIMESERIES <Path> WHERE TAGS(key)='value'` or `COUNT TIMESERIES <Path> WHERE TAGS(key) contains 'value'`.
* `LEVEL` could be defined to show count the number of timeseries of each node at the given level in current Metadata Tree. This could be used to query the number of sensors under each device. The grammar is: `COUNT TIMESERIES <Path> GROUP BY LEVEL=<INTEGER>`.


```
IoTDB > COUNT TIMESERIES root.**
IoTDB > COUNT TIMESERIES root.ln.**
IoTDB > COUNT TIMESERIES root.ln.*.*.status
IoTDB > COUNT TIMESERIES root.ln.wf01.wt01.status
IoTDB > COUNT TIMESERIES root.** WHERE TIMESERIES contains 'sgcc' 
IoTDB > COUNT TIMESERIES root.** WHERE DATATYPE = INT64
IoTDB > COUNT TIMESERIES root.** WHERE TAGS(unit) contains 'c' 
IoTDB > COUNT TIMESERIES root.** WHERE TAGS(unit) = 'c' 
IoTDB > COUNT TIMESERIES root.** WHERE TIMESERIES contains 'sgcc' group by level = 1
```

For example, if there are several timeseries (use `show timeseries` to show all timeseries):

```
+-------------------------------+--------+-------------+--------+--------+-----------+-------------------------------------------+--------------------------------------------------------+--------+-------------------+
|                     timeseries|   alias|     database|dataType|encoding|compression|                                       tags|                                              attributes|deadband|deadband parameters|
+-------------------------------+--------+-------------+--------+--------+-----------+-------------------------------------------+--------------------------------------------------------+--------+-------------------+
|root.sgcc.wf03.wt01.temperature|    null|    root.sgcc|   FLOAT|     RLE|     SNAPPY|                                       null|                                                    null|    null|               null|
|     root.sgcc.wf03.wt01.status|    null|    root.sgcc| BOOLEAN|   PLAIN|     SNAPPY|                                       null|                                                    null|    null|               null|
|             root.turbine.d1.s1|newAlias| root.turbine|   FLOAT|     RLE|     SNAPPY|{"newTag1":"newV1","tag4":"v4","tag3":"v3"}|{"attr2":"v2","attr1":"newV1","attr4":"v4","attr3":"v3"}|    null|               null|
|     root.ln.wf02.wt02.hardware|    null|      root.ln|    TEXT|   PLAIN|     SNAPPY|                               {"unit":"c"}|                                                    null|    null|               null|
|       root.ln.wf02.wt02.status|    null|      root.ln| BOOLEAN|   PLAIN|     SNAPPY|                    {"description":"test1"}|                                                    null|    null|               null|
|  root.ln.wf01.wt01.temperature|    null|      root.ln|   FLOAT|     RLE|     SNAPPY|                                       null|                                                    null|    null|               null|
|       root.ln.wf01.wt01.status|    null|      root.ln| BOOLEAN|   PLAIN|     SNAPPY|                                       null|                                                    null|    null|               null|
+-------------------------------+--------+-------------+--------+--------+-----------+-------------------------------------------+--------------------------------------------------------+--------+-------------------+
Total line number = 7
It costs 0.004s
```

Then the Metadata Tree will be as below:

<center><img style="width:100%; max-width:600px; margin-left:auto; margin-right:auto; display:block;" src="https://alioss.timecho.com/docs/img/github/69792176-1718f400-1201-11ea-861a-1a83c07ca144.jpg"></center>
As can be seen, `root` is considered as `LEVEL=0`. So when you enter statements such as:

```
IoTDB > COUNT TIMESERIES root.** GROUP BY LEVEL=1
IoTDB > COUNT TIMESERIES root.ln.** GROUP BY LEVEL=2
IoTDB > COUNT TIMESERIES root.ln.wf01.* GROUP BY LEVEL=2
```

You will get following results:

```
+------------+-----------------+
|      column|count(timeseries)|
+------------+-----------------+
|   root.sgcc|                2|
|root.turbine|                1|
|     root.ln|                4|
+------------+-----------------+
Total line number = 3
It costs 0.002s

+------------+-----------------+
|      column|count(timeseries)|
+------------+-----------------+
|root.ln.wf02|                2|
|root.ln.wf01|                2|
+------------+-----------------+
Total line number = 2
It costs 0.002s

+------------+-----------------+
|      column|count(timeseries)|
+------------+-----------------+
|root.ln.wf01|                2|
+------------+-----------------+
Total line number = 1
It costs 0.002s
```

> Note: The path of timeseries is just a filter condition, which has no relationship with the definition of level.

## Tag and Attribute Management

We can also add an alias, extra tag and attribute information while creating one timeseries.

The differences between tag and attribute are:

* Tag could be used to query the path of timeseries, we will maintain an inverted index in memory on the tag: Tag -> Timeseries
* Attribute could only be queried by timeseries path : Timeseries -> Attribute

The SQL statements for creating timeseries with extra tag and attribute information are extended as follows:

```
create timeseries root.turbine.d1.s1(temprature) with datatype=FLOAT, encoding=RLE, compression=SNAPPY tags(tag1=v1, tag2=v2) attributes(attr1=v1, attr2=v2)
```

The `temprature` in the brackets is an alias for the sensor `s1`. So we can use `temprature` to replace `s1` anywhere.

> IoTDB also supports [using AS function](../Reference/SQL-Reference.md#data-management-statement) to set alias. The difference between the two is: the alias set by the AS function is used to replace the whole time series name, temporary and not bound with the time series; while the alias mentioned above is only used as the alias of the sensor, which is bound with it and can be used equivalent to the original sensor name.

> Notice that the size of the extra tag and attribute information shouldn't exceed the `tag_attribute_total_size`.

We can update the tag information after creating it as following:

* Rename the tag/attribute key
```
ALTER timeseries root.turbine.d1.s1 RENAME tag1 TO newTag1
```
* Reset the tag/attribute value
```
ALTER timeseries root.turbine.d1.s1 SET newTag1=newV1, attr1=newV1
```
* Delete the existing tag/attribute
```
ALTER timeseries root.turbine.d1.s1 DROP tag1, tag2
```
* Add new tags
```
ALTER timeseries root.turbine.d1.s1 ADD TAGS tag3=v3, tag4=v4
```
* Add new attributes
```
ALTER timeseries root.turbine.d1.s1 ADD ATTRIBUTES attr3=v3, attr4=v4
```
* Upsert alias, tags and attributes
> add alias or a new key-value if the alias or key doesn't exist, otherwise, update the old one with new value.
```
ALTER timeseries root.turbine.d1.s1 UPSERT ALIAS=newAlias TAGS(tag3=v3, tag4=v4) ATTRIBUTES(attr3=v3, attr4=v4)
```
* Show timeseries using tags. Use TAGS(tagKey) to identify the tags used as filter key
```
SHOW TIMESERIES (<`PathPattern`>)? timeseriesWhereClause
```

returns all the timeseries information that satisfy the where condition and match the pathPattern. SQL statements are as follows:

```
ALTER timeseries root.ln.wf02.wt02.hardware ADD TAGS unit=c
ALTER timeseries root.ln.wf02.wt02.status ADD TAGS description=test1
show timeseries root.ln.** where TAGS(unit)='c'
show timeseries root.ln.** where TAGS(description) contains 'test1'
```

The results are shown below respectly:

```
+--------------------------+-----+-------------+--------+--------+-----------+------------+----------+--------+-------------------+
|                timeseries|alias|     database|dataType|encoding|compression|        tags|attributes|deadband|deadband parameters|
+--------------------------+-----+-------------+--------+--------+-----------+------------+----------+--------+-------------------+
|root.ln.wf02.wt02.hardware| null|      root.ln|    TEXT|   PLAIN|     SNAPPY|{"unit":"c"}|      null|    null|               null|
+--------------------------+-----+-------------+--------+--------+-----------+------------+----------+--------+-------------------+
Total line number = 1
It costs 0.005s

+------------------------+-----+-------------+--------+--------+-----------+-----------------------+----------+--------+-------------------+
|              timeseries|alias|     database|dataType|encoding|compression|                   tags|attributes|deadband|deadband parameters|
+------------------------+-----+-------------+--------+--------+-----------+-----------------------+----------+--------+-------------------+
|root.ln.wf02.wt02.status| null|      root.ln| BOOLEAN|   PLAIN|     SNAPPY|{"description":"test1"}|      null|    null|               null|
+------------------------+-----+-------------+--------+--------+-----------+-----------------------+----------+--------+-------------------+
Total line number = 1
It costs 0.004s
```

- count timeseries using tags

```
COUNT TIMESERIES (<`PathPattern`>)? timeseriesWhereClause
COUNT TIMESERIES (<`PathPattern`>)? timeseriesWhereClause GROUP BY LEVEL=<INTEGER>
```

returns all the number of timeseries that satisfy the where condition and match the pathPattern. SQL statements are as follows:

```
count timeseries
count timeseries root.** where TAGS(unit)='c'
count timeseries root.** where TAGS(unit)='c' group by level = 2
```

The results are shown below respectly :

```
IoTDB> count timeseries
+-----------------+
|count(timeseries)|
+-----------------+
|                6|
+-----------------+
Total line number = 1
It costs 0.019s
IoTDB> count timeseries root.** where TAGS(unit)='c'
+-----------------+
|count(timeseries)|
+-----------------+
|                2|
+-----------------+
Total line number = 1
It costs 0.020s
IoTDB> count timeseries root.** where TAGS(unit)='c' group by level = 2
+--------------+-----------------+
|        column|count(timeseries)|
+--------------+-----------------+
|  root.ln.wf02|                2|
|  root.ln.wf01|                0|
|root.sgcc.wf03|                0|
+--------------+-----------------+
Total line number = 3
It costs 0.011s
```

> Notice that, we only support one condition in the where clause. Either it's an equal filter or it is an `contains` filter. In both case, the property in the where condition must be a tag.

create aligned timeseries

```
create aligned timeseries root.sg1.d1(s1 INT32 tags(tag1=v1, tag2=v2) attributes(attr1=v1, attr2=v2), s2 DOUBLE tags(tag3=v3, tag4=v4) attributes(attr3=v3, attr4=v4))
```

The execution result is as follows:

```
IoTDB> show timeseries
+--------------+-----+-------------+--------+--------+-----------+-------------------------+---------------------------+--------+-------------------+
|    timeseries|alias|     database|dataType|encoding|compression|                     tags|                 attributes|deadband|deadband parameters|
+--------------+-----+-------------+--------+--------+-----------+-------------------------+---------------------------+--------+-------------------+
|root.sg1.d1.s1| null|     root.sg1|   INT32|     RLE|     SNAPPY|{"tag1":"v1","tag2":"v2"}|{"attr2":"v2","attr1":"v1"}|    null|               null|
|root.sg1.d1.s2| null|     root.sg1|  DOUBLE| GORILLA|     SNAPPY|{"tag4":"v4","tag3":"v3"}|{"attr4":"v4","attr3":"v3"}|    null|               null|
+--------------+-----+-------------+--------+--------+-----------+-------------------------+---------------------------+--------+-------------------+
```

Support query：

```
IoTDB> show timeseries where TAGS(tag1)='v1'
+--------------+-----+-------------+--------+--------+-----------+-------------------------+---------------------------+--------+-------------------+
|    timeseries|alias|     database|dataType|encoding|compression|                     tags|                 attributes|deadband|deadband parameters|
+--------------+-----+-------------+--------+--------+-----------+-------------------------+---------------------------+--------+-------------------+
|root.sg1.d1.s1| null|     root.sg1|   INT32|     RLE|     SNAPPY|{"tag1":"v1","tag2":"v2"}|{"attr2":"v2","attr1":"v1"}|    null|               null|
+--------------+-----+-------------+--------+--------+-----------+-------------------------+---------------------------+--------+-------------------+
```

The above operations are supported for timeseries tag, attribute updates, etc.