Flux Query Language中SELECT指定列的等价语法及示例查询咨询

Great question—this is definitely one of those areas where Flux's functional approach can feel a bit disjointed from SQL's familiar structure, especially since the docs focus heavily on filter() for the WHERE-equivalent logic and gloss over column selection.

Let's break this down clearly:

In SQL, SELECT <columns> lets you pick specific fields to return from your dataset. In Flux, there's no direct 1:1 "SELECT" keyword—instead, you use data transformation functions to shape your output. The most straightforward equivalent for picking specific columns is the keep() function, which retains only the columns you specify and drops everything else.

For your exact example: SELECT address, name FROM addresses #

Assuming your "addresses" data is stored as a measurement named addresses in your InfluxDB bucket, the equivalent Flux query would look like this:

// First, pull the base dataset from your bucket
from(bucket: "your-bucket-name")
  // Define the time range (use -inf/+inf for all time, adjust as needed)
  |> range(start: -inf, stop: +inf)
  // Filter to only the "addresses" measurement (equivalent to FROM addresses)
  |> filter(fn: (r) => r._measurement == "addresses")
  // Keep only the address and name columns (equivalent to SELECT address, name)
  |> keep(columns: ["address", "name"])

Alternative: Using map() to restructure rows #

If you want more control over the output shape (like renaming columns or calculating values on the fly), you can use map() to explicitly define the fields you want to keep:

from(bucket: "your-bucket-name")
  |> range(start: -inf, stop: +inf)
  |> filter(fn: (r) => r._measurement == "addresses")
  |> map(fn: (r) => ({
    address: r.address,
    name: r.name
  }))

A quick note on the docs #

You’re right that the existing docs mention filter() as equivalent to SELECT + WHERE, but that’s a bit misleading. Really, filter() handles only the row-filtering logic (the WHERE part). The column selection part (the SELECT columns) is handled separately with functions like keep() or map(). This separation is because Flux is built around a pipeline of transformations—each step modifies the data incrementally rather than declaring the entire output in one clause.

内容的提问来源于stack exchange，提问作者Sushovan Mandal