Skip to main content
Version: v1.9.0



from source data from pools, files, or URIs


from <pool>[@<commitish>]
from <pattern>
file <path> [format <format>]
get <uri> [format <format>]
from (
pool <pool>[@<commitish>] [ => <leg> ]
pool <pattern>
file <path> [format <format>] [ => <leg> ]
get <uri> [format <format>] [ => <leg> ]


The from operator identifies one or more data sources and transmits their data to its output. A data source can be

  • the name of a data pool in a Zed lake, with optional commitish;
  • the names of multiple data pools, expressed as a regular expression or glob pattern;
  • a path to a file;
  • an HTTP, HTTPS, or S3 URI; or
  • the pass operator, to treat the upstream data path as a source.

File paths and URIs may be followed by an optional format specifier.

Sourcing data from pools is only possible when querying a lake, such as via the zed command or Zed lake API. Sourcing data from files is only possible with the zq command.

When a single pool name is specified without @-referencing a commit or ID, or when using a pool pattern, the tip of the main branch of each pool is accessed.

In the first four forms, a single source is connected to a single output. In the fifth form, multiple sources are accessed in parallel and may be joined, combined, or merged.

A data path can be split with the fork operator as in

from PoolOne | fork (
=> op1 | op2 | ...
=> op1 | op2 | ...
) | merge ts | ...

Or multiple pools can be accessed and, for example, joined:

from (
pool PoolOne => op1 | op2 | ...
pool PoolTwo => op1 | op2 | ...
) | join on key=key | ...

Similarly, data can be routed to different paths with replication using the switch operator:

from ... | switch color (
case "red" => op1 | op2 | ...
case "blue" => op1 | op2 | ...
default => op1 | op2 | ...
) | ...

Input Data

Examples below below assume the existence of the Zed lake created and populated by the following commands:

export ZED_LAKE=example
zed -q init
zed -q create -orderby flip:desc coinflips
echo '{flip:1,result:"heads"} {flip:2,result:"tails"}' | zed load -q -use coinflips -
zed branch -q -use coinflips trial
echo '{flip:3,result:"heads"}' | zed load -q -use coinflips@trial -
zed -q create numbers
echo '{number:1,word:"one"} {number:2,word:"two"} {number:3,word:"three"}' | zed load -q -use numbers -
zed query -f text 'from :branches | yield + "@" + | sort'

The lake then contains the two pools:


The following file hello.zson is also used.

{greeting:"hello world!"}


Source structured data from a local file

zq -z 'file hello.zson | yield greeting'


"hello world!"

Source data from a local file, but in line format

zq -z 'file hello.zson format line'


"{greeting:\"hello world!\"}"

Source structured data from a URI

zq -z 'get | yield productName'


"Zui - Insiders"

Source data from the main branch of a pool

zed -lake example query -z 'from coinflips'



Source data from a specific branch of a pool

zed -lake example query -z 'from coinflips@trial'



Count the number of values in the main branch of all pools

zed -lake example query -f text 'from * | count()'



Join the data from multiple pools

zed -lake example query -z '
from coinflips | sort flip
| join (
from numbers | sort number
) on flip=number word'



Use pass to combine our join output with data from yet another source

zed -lake example query -z '
from coinflips | sort flip
| join (
from numbers | sort number
) on flip=number word
| from (
pool coinflips@trial => c:=count() | yield "There were ${int64(c)} flips"
) | sort this'


"There were 3 flips"