Skip to contents

Load and row-bind partitioned data

Source: R/partitions.R
st_load_parts.Rd

Load and row-bind partitioned data

Usage

st_load_parts(base, filter = NULL, columns = NULL, as = c("rbind", "dt"))

Arguments

base

Base dir

filter

Partition filter. Supports three formats:

  • Named list for exact matching: list(country = "USA", year = 2020)

  • Formula with expression: ~ year > 2010 or ~ country == "COL" & year >= 2012

  • NULL for no filtering (default)

columns

Character vector of column names to load (optional). For parquet/fst formats, uses native column selection (fast, low memory). For other formats (qs/rds/csv), loads full object then subsets (with warning).

as

Data frame binding mode: "rbind" (base) or "dt" (data.table)

Value

Data frame with unioned columns and extra columns for the key fields

Examples

if (FALSE) { # \dontrun{
# Load all partitions
st_load_parts("data/parts")

# Filter with exact match
st_load_parts("data/parts", filter = list(country = "USA"))

# Filter with expression
st_load_parts("data/parts", filter = ~ year > 2010)

# Combine filter + column selection
st_load_parts("data/parts", filter = ~ year > 2010, columns = c("value", "metric"))
} # }