infludb语法–官网 features a SQL like query language, only used for querying data. The HTTP API has endpoints for writing data and performing other database administration tasks. The only

InfluxDB features a SQL like query language,only used for querying data. The HTTP API has endpoints for writing data and performing other database administration tasks. The only exception to this is,which perpetually write their results into one or more time series.

InfluxDB allows you to use any characters in your time series names. However,parsing queries for those series can be tricky. So it’s best to wrap your queries for any series that has characters other than letters in double quotes like this:

select * from "series with \"double quotes\""

<h2 id=””>
<a id=”getting-a-list-of-time-series” class=”offset-anchor”><a href=””&gt;Getting a List of Time Series

There are two different methods for returning a list of all time series in a database:

The first query will return all series,while the second will return the most recent point from each series that matches the given regex.

By default,InfluxDB returns data in time descending order. The most efficient queries run over only a single column in a given time series.

This simple query pulls the values for thevaluecolumn from theresponse_timesseries.

If start and end times aren’t set they will default to beginning of time until now,respectively.

The columntimeis built in for every time series in the database. You specify the start and end times by setting conditions on thetimecolumns in the where clause.

Below are the different formats that can be used to specify start and end times.

Date time strings have the formatYYYY-MM-DD HH:MM:SS.mmmwheremmmare the milliseconds within the second. For example:

 '2013-08-12 23:32:01.232' and time < '2013-08-13';

The time and date should be wrapped in single quotes. If you only specify the date,the time will be set to00:00:00. The.232after the hours,minutes,and seconds is optional and specifies the milliseconds.

You can usenow()to calculate a timestamp relative to the server’s current timestamp. For example:

 now() - 1h limit 1000;

will return all points starting an hour ago until now.

Other options for how to specify time durations areufor microseconds,sfor seconds,mfor minutes,hfor hours,dfor days andwfor weeks. If no suffix is given the value is interpreted as microseconds.

You can specify timestamp in epoch time,which is defined as the number of microseconds that have elapsed since 00:00:00 Coordinated Universal Time (UTC),Thursday,1 January 1970. You can use the same suffixes from the previous section if you don’t want to specify timestamp in microseconds. For example:


will return all points that were writtern after2014-01-01 00:00:00

Points are uniquely identified by the time series they appear in,the time,and the sequence number. Here’s a query that returns a specific point.

Notethat the time is a very large number. That’s because it’s a microsecond scale epoch. InfluxDB always stores points at this scale,but most libraries will return the time as either a second,or millisecond scale value. If you’re selecting a specific point,you’ll need to know the exact microsecond scale epoch that point has otherwise you’ll get an unexpected empty result.

You can select from multiple series by name or by specifying a regex to match against. Here are a few examples.

Get the last hour of data from the two seriesevents,anderrors. Here’s a regex example:

 now() - 1h;

Get the last hour of data from every time series that starts withstats.(case insensitive). Another example:

Return the last point from every time series in the database.

The delete query looks like the following:

With no time constraints this query will delete every point in the time seriesresponse_times. You must be a cluster or database admin to run delete queries.

You can also delete from any series that matches a regex:

Any conditions in the where clause that don’t set the start and/or end time will be ignored,for example the following query returns an error:

Delete time conditions only support ranges,an equals condition (=) is currently not supported.

Deleting all data for a series will only remove the points. It will still remain in the index. If you want to remove all data from a series and remove it from the list of series in a database use thedropquery:

Currently,deletes are not very efficient. If you want to quickly evict old data,the best way to do that is by dropping a shard. For more.

We’ve already seen the where clause for selecting time ranges and a specific point. You can also use it to filter based on given values,comparators,or regexes. Here are some examples of different ways to usewhere.

select * from log_lines where line =~ /error/i;

select * from events where customer_id = 23 and type = 'click';

select * from response_times where value > 500;

select from events where email !~ /.gmail.*/;

select * from nagios_checks where status <> 0;

select * from events where signed_in = false;

select from events
where (email =~ /.
gmail./ or email =~ /.yahoo.*/) and state = 'ny';

The where clause supports comparisons against regexes,strings,booleans,floats,integers,and the times listed before. Comparators include=equal to,>greater than,<less than,<>not equal to,=~matches against,!~doesn’t match against. You can chain logic together usingandandorand you can separate using(and)

The group by clause in InfluxDB is used not only for grouping by given values,but also for grouping by given time buckets. You’ll always be pairing this up within theselectclause. Here are a few examples to illustrate how group by works.

-- count of each unique type of event in 10 minute intervals
select count(type) from events group by time(10m),type;

-- 95th percentile of response times in 30 second intervals
select percentile(value,95) from response_times group by time(30s);

作者: dawei






在线咨询: QQ交谈