Extract.pipeflow_pip: Extract or subset a pipeline

In pipeflow: Fast Interactive Data Analysis Pipelines

Extract or subset a pipeline

Description

Returns a new pipeline containing selected steps and all required upstream dependencies.

Extracts values from a pipeline using one or two indices. With a single string name, named fields such as "pipeline" or "name" are returned first; anything else returns the matching step-table column. With two indices (row, column), a single cell is extracted.

Usage

## S3 method for class 'pipeflow_pip'
x[i, ...]

## S3 method for class 'pipeflow_pip'
x[[i, j, ...]]

Arguments

x	A pipeflow pipeline object.
i	integer (row indices) or character vector (step names) of steps to select
...	not used
j	column names to select

Value

A new pipeflow pipeline object.

Extracted value(s), depending on i and j.

Examples

p <- pip_new() |>
  pip_add("load", \(n = 5) seq_len(n)) |>
  pip_add("square", \(x = ~load) x^2) |>
  pip_add("total", \(x = ~square) sum(x))

# Select by step name — upstream deps are pulled in automatically.
# Selecting only "total" still includes "load" and "square".
sub <- p["total"]
sub[["pipeline"]][["step"]] # "load", "square", "total"

# Select a subset of steps by name vector
p[c("load", "square")][["pipeline"]][["step"]] # "load", "square"

# Select by integer row index
p[1:2][["pipeline"]][["step"]] # "load", "square"
p <- pip_new() |>
  pip_add("load", \(x = 1) x) |>
  pip_add("fit", \(x = ~load) x + 1)

# Access internal objects by name
p[["pipeline"]] # the full step table
p[["name"]] # "pipe"

# Shorthand column access (equivalent to p[["pipeline"]][["step"]])
p[["step"]]

# Two-index form: p[[row, column]] extracts a single cell
p[["fit", "depends"]] # "load"
p[[2, "state"]] # state of the second step

pipeflow documentation built on June 15, 2026, 9:10 a.m.