Doubts about using the incluxdb database

InfluxDB Is an open source database for time series data storage , Here are some doubts in the process of use （ in the light of 1.8.x edition ）.

1. About Client edition
   The new version is recommended in the official documents Python Client,Python Package name is influxdb_client, This client is for InfluxDB2.0 Version developed , But compatible 1.8.x edition . If there is a new version, there must be an old version ,Python Package name is influxdb, The adaptation is mentioned in its documentation 1.7 Up to . Actual test findings , The old client can also be in 1.8.1 Upper use .

2. About query language
   There are currently two query languages , They are classic models InfluxQL And new models Flux. The two are different in grammar , There is no in-depth understanding of , So I don't know the advantages and disadvantages .
   2.1 client
   It's worth make complaints about. , The old version client uses InfluxQL Language , The new version client uses Flux. The query interfaces provided are also different , There are great differences in the organization and format of the returned results ：

   • InfluxQL Provide ResultSet and DataFrame Query interface .
   • Flux Provide FluxTable、CSV、DataFrame and Raw Data Four kinds of queries API.

   2.2 Date and time

   • InfluxQL Support RFC3339 Format and class RFC3339 Format
   • Flux Support only RFC3339 Format , The time part is optional , If the time is omitted, it defaults to midnight of the day ( zero ).

   2.3 HTTP API
   Two query languages HTTP API It's a little different , There are great differences in the organization and format of the returned results ：

   • InfluxQL Default GET request , You can also use POST request , return json Results in format .
   • Flux Accept POST request , return annotated CSV Results in format .
     Please check the document for details ：Query data with the InfluxDB API

3. About Flux Query returns by default DataFrame list
   In the use of influxdb_client Of query_data_frame Query the same measurement when , By default, the returned result is always a DataFrame list ,g The root cause is that it defaults to tag,field,measurement Conduct group. If you want to return a table , have access to |> group() Mask default group operation .

4. About influxdb in InfluxDBClient and DataFrameClient The difference between the return result format of the interface
   InfluxDBClient：
   Single sentence query returns ResultSet
   Multiple queries return list, Element is ResultSet
   DataFrameClient：
   Single SHOW sentence , The type of the returned result is ResultSet
   Multiple SHOW sentence , The type of the returned result is list, Element is ResultSet
   Single SELECT sentence , The return result is collections.defaultdict, Its key Is the name of the watch ,value by DataFrame.
   Multiple SELECT sentence , The return result is map iterator （ Yes Python3 for ）, The iterative value is collections.defaultdict.
   To sum up, combined with the source code, we can get ： only SELECT The first sentence will turn into DataFrame Format and the format of the returned result is different .

5. influx and influxd
influx Start a command line interface , Executable for management InfluxDB Many aspects of the order .
   influxd Start and run InfluxDB Processes required for normal operation .
   for example ,influxd config Default display configuration

6. About hardware configuration requirements
   Used in projects InfluxDB It is recommended to evaluate the hardware configuration requirements first , And in use , Adjust in time according to the situation .Hardware sizing guidelines | InfluxDB OSS 1.8 Documentation

7. Key concepts
   point： Represents a data record , from measurement、tag set、field set as well as timestamp form . By its series and timestamp Unique identification .
   series： from measurement、tag set、field key Logical grouping of defined data .