Flux query basics
Most Flux queries follow the same basic structure. Familiarize yourself with the basic concepts and steps to use when querying data with Flux.
Basic query structure
The majority of basic Flux queries include the following steps:
from(bucket: "example-bucket") // ── Source |> range(start: -1d) // ── Filter on time |> filter(fn: (r) => r._field == "foo") // ── Filter on column values |> group(columns: ["sensorID"]) // ── Shape |> mean() // ── Process
Flux supports multiple data sources including, time series databases (such as InfluxDB and Prometheus), relational databases (such as MySQL and PostgreSQL), CSV, and more.
- For more information about supported data sources, see Query data sources.
- For a list of input functions, see Function types and categories – Inputs.
Filter functions iterate over and evaluate each input row to see if it matches specified conditions. Rows that meet the conditions are included in the function output. Rows that do not meet the specified conditions are dropped.
Flux provides the following primary filter functions:
range(): filter data based on time.
filter(): filter data based on column values.
filter()uses a predicate function defined in the
fnparameter to evaluate input rows. Each row is passed into the predicate function as a record,
r, containing key-value pairs for each column in the row.
Other filter functions are also available. For more information, see Function types and categories – Filters.
Many queries need to modify the structure of data to prepare it for processing. Common data-shaping tasks include regrouping data by column values or by time or pivoting column values into rows.
Functions that reshape data include the following:
group(): modify group keys
_stopvalues of rows to group data by time
pivot(): pivot column values into rows
drop(): drop specific columns
keep(): keep specific columns and drop all others
Processing data can take on many forms, and includes the following types of operations:
- Aggregate data: aggregate all rows of an input table into a single row. For information, see Function types and categories - Aggregates.
- Select specific data points: return specific rows from each input table. For example, return the first or last row, the row with the highest or lowest value, and more. For information, see Function types and categories - Selectors.
- Rewrite rows: use
map()to rewrite each input row. Tranform values with mathematic operations, process strings, dynamically add new columns, and more.
- Send notifications: evaluate data and use Flux notification endpoint functions to send notifications to external services. For information, see Function types and categories- Notification endpoints.
aggregateWindow helper function
aggregateWindow() is a helper function
that both shapes and processes data.
The function windows and groups data by time, and then applies an aggregate
or selector function to the restructured tables.
Write a basic query
import "influxdata/influxdb/sample" sample.data(set: "airSensor")
sample.data()returns data as if it was queried from InfluxDB. To actually query data from InfluxDB, replace
Pipe the returned data forward into
range()to filter the data by time. Return data from the last hour.
import "influxdata/influxdb/sample" sample.data(set: "airSensor") |> range(start: -1h)
filter()to filter rows based on column values. In this example, return only rows that include values for the
cofield. The field name is stored in the
import "influxdata/influxdb/sample" sample.data(set: "airSensor") |> range(start: -1h) |> filter(fn: (r) => r._field == "co")
mean()to calculate the average value in each input table. Because InfluxDB groups data by series,
mean()returns a table for each unique
sensor_idcontaining a single row with the average value in the
import "influxdata/influxdb/sample" sample.data(set: "airSensor") |> range(start: -1h) |> filter(fn: (r) => r._field == "co") |> mean()
import "influxdata/influxdb/sample" sample.data(set: "airSensor") |> range(start: -1h) |> filter(fn: (r) => r._field == "co") |> mean() |> group()
Results from this basic query should be similar to the following:
_stop columns have been omitted.
Support and feedback
Thank you for being part of our community! We welcome and encourage your feedback and bug reports for Flux and this documentation. To find support, the following resources are available:
InfluxDB Cloud customers can contact InfluxData Support.