Command types

There are six broad types for all of the search commands: distributable streaming, centralized streaming, transforming, generating, orchestrating and dataset processing. These types are not mutually exclusive. A command might be streaming or transforming, and also generating.

The following tables list the commands that fit into each of these types. For detailed explanations about each of the types, see Types of commands in the Search Manual.

To find out how the types of commands used in searches can affect performance, see Write better searches in the Search Manual.

Streaming commands

A streaming command operates on each event as the event is returned by a search.

• A distributable streaming command runs on the indexer or the search head, depending on where in the search the command is invoked. Any distributable streaming command that comes after a non-streaming command in the search is processed on the search head.
• A centralized streaming command applies a transformation to each event returned by a search. Unlike distributable streaming commands, a centralized streaming command only works on the search head.

Command	Notes
addinfo	Distributable streaming
addtotals	Distributable streaming. A transforming command when used to calculate column totals (not row totals).
arules	Some of the work is distributable streaming running on the indexer or the search head. The rest of the work is centralized streaming running on the search head.
autoregress	Centralized streaming.
bin	Streaming if specified with the span argument. Otherwise a dataset processing command.
bucketdir	Distributable streaming by default, but centralized streaming if the local setting specified for the command in the commands.conf file is set to true.
cluster	Streaming in some modes.
convert	Distributable streaming.
dedup	Distributable streaming in a prededup phase. Centralized streaming after the individual indexers perform their own dedup and the results are returned to the search head from each indexer. Using the sortby argument or specifying keepevents=true makes the dedup command a dataset processing command.
eval	Distributable streaming.
extract	Distributable streaming.
fieldformat	Distributable streaming.
fields	Distributable streaming.
fillnull	Distributable streaming when a field-list is specified. A dataset processing command when no field-list is specified.
head	Centralized streaming.
highlight	Distributable streaming.
iconify	Distributable streaming.
iplocation	Distributable streaming.
join	Centralized streaming, if there is a defined set of fields to join to. A dataset processing command when no field-list is specified.
lookup	Distributable streaming when specified with local=false, which is the default. An orchestrating command when local=true.
makemv	Distributable streaming.
multikv	Distributable streaming.
mvexpand	Distributable streaming.
nomv	Distributable streaming.
rangemap	Distributable streaming.
regex	Distributable streaming.
reltime	Distributable streaming.
rename	Distributable streaming.
replace	Distributable streaming.
rex	Distributable streaming.
search	Distributable streaming if used further down the search pipeline. A generating command when it is the first command in the search.
spath	Distributable streaming.
strcat	Distributable streaming.
streamstats	Centralized streaming.
tags	Distributable streaming.
transaction	Centralized streaming.
typer	Distributable streaming.
where	Distributable streaming.
untable	Distributable streaming.
xmlkv	Distributable streaming.
xmlunescape	Distributable streaming by default, but centralized streaming if the local setting specified for the command in the commands.conf file is set to true.
xpath	Distributable streaming.
xyseries	Distributable streaming if the argument grouped=false is specified, which is the default. Otherwise a transforming command.

Generating commands

A generating command either returns information or generates results. Some generating commands can return information from an index, a data model, a lookup, or a CSV file without any transformations to the information. Other generating commands generate results, usually for testing purposes.

Command	Notes
datamodel	Report-generating
dbinspect	Report-generating.
eventcount	Report-generating.
from	Can be either report-generating or event-generating depending on the search or knowledge object that is referenced by the command.
gentimes	Event-generating.
inputcsv	Event-generating (centralized).
Inputlookup	Event-generating (centralized) when append=false, which is the default.
loadjob	Event-generating (centralized).
makeresults	Report-generating.
metadata	Report-generating. Although metadata fetches data from all peers, any command run after it runs only on the search head.
metasearch	Event-generating.
mstats	Report-generating, except when append=true is specified.
multisearch	Event-generating.
pivot	Report-generating.
rest
search	Event-generating (distributable) when the first command in the search, which is the default. A streaming (distributable) command if used later in the search pipeline.
searchtxn	Event-generating.
set	Event-generating.
tstats	Report-generating (distributable), except when prestats=true. When prestats=true, the tstats command is event-generating.

Transforming commands

A transforming command orders the results into a data table. The command "transforms" the specified cell values for each event into numerical values for statistical purposes.

In earlier versions of Splunk software, transforming commands were called reporting commands.

Command	Notes
addtotals	Transforming when used to calculate column totals (not row totals). A distributable streaming command when used to calculate row totals, which is the default.
anomalydetection
append
associate
chart
cofilter
contingency
history
makecontinuous
mvcombine
rare
stats
table
timechart
top
xyseries	Transforming if grouped=true. A streaming (distributable) command when grouped=false, which is the default setting.

Orchestrating commands

Orchestrating commands control some aspect of how a search is processed. They do not directly affect the final result set of the search. For example, you might apply an orchestrating command to a search to enable or disable a search optimization that helps the overall search complete faster.

Command	Notes
localop
lookup	Only becomes an orchestrating command when local=true. This forces the lookup command to run on the search head and not on any remote peers. A streaming (distributable) command when local=false, which is the default setting.
noop
redistribute
require

Dataset processing commands

A dataset processing command is a command that requires the entire dataset before the command can run. Some of these commands fit into other command types in specific situations or when specific arguments are used.

Command	Notes
anomalousvalue	Some modes
anomalydetection	Some modes
append	Some modes
appendcols
appendpipe
bin	Some modes. A streaming command if the span argument is specified.
cluster	Some modes
concurrency
datamodel
dedup	Using the sortby argument or specifying keepevents=true makes the dedup command a dataset processing command. Otherwise, dedup is a distributable streaming command in a prededup phase. Centralized streaming after the individual indexers perform their own dedup and the results are returned to the search head from each indexer.
eventstats
fieldsummary
fillnull	When no field-list is specified, a dataset processing command. If a field-list is specified fillnull is a distributable streaming command.
from	Some modes
join	Some modes. A centralized streaming command when there is a defined set of fields to join to.
map
outlier
reverse
sort
tail
transaction	Some modes
union	Some modes