Skip to contents

GeoParquet Row Groups and Statistics

Source: R/gpq_meta.R
gpq_row_groups.Rd

Read row group and column chunk statistics. Parquet stores chunk min/max statistics as raw bytes in the column's physical type; these are decoded to R scalars and aggregated into a per-column summary across all row groups.

Usage

gpq_row_groups(gpq_path)

Arguments

gpq_path

Path to a (Geo)Parquet (*.parquet) file.

Value

A gpq_row_groups object (a list) with:

  • source, path: the file name and full path.

  • row_groups: a per-row-group tibble::tibble().

  • column_chunks: the raw per-chunk tibble::tibble() with leaf column names and decoded min/max appended.

  • col_summary: a one-row-per-column tibble::tibble() aggregated across all row groups.

Examples

f <- system.file("extdata/atlanta/atlanta.parquet", package = "gdalvector")
gpq_row_groups(f)
#> <gpq_row_groups/list> (1 row group)
#> # A tibble: 6 × 10
#>   column col_name    type     codec num_values null_count min   max   compressed
#>    <int> <chr>       <chr>    <chr>      <dbl>      <dbl> <chr> <chr>      <byt>
#> 1      0 OGC_FID     INT64    ZSTD       44578          0 0     44,5…   46.02 kB
#> 2      1 source_fid  BYTE_AR… ZSTD       44578          0 0001… fffd…  930.22 kB
#> 3      2 geoid       BYTE_AR… ZSTD       44578          0 13089 13121      254 B
#> 4      3 state_fips  BYTE_AR… ZSTD       44578          0 13    13          69 B
#> 5      4 county_fips BYTE_AR… ZSTD       44578          0 089   121        250 B
#> 6      5 geom        BYTE_AR… ZSTD       44578          0 -     -        1.36 MB
#> # ℹ 1 more variable: uncompressed <byt>