lazysf {lazysf}

Delayed (lazy) read for GDAL vector

Description

A lazy data frame for GDAL drawings ('vector data sources'). lazysf is DBI compatible and designed to work with dplyr. It should work with any data source (file, url, connection string) readable by the sf package function sf_read.

Usage

lazysf(x, layer, ...)

## S3 method for class 'character'
lazysf(x, layer, ..., query = NA)

## S3 method for class 'SFSQLConnection'
lazysf(x, layer, ..., query = NA)

Arguments

x

the data source name (file path, url, or database connection string

• analogous to sf::read_sf() 'dsn')

layer

layer name (varies by driver, may be a file name without extension); in case layer is missing, st_read will read the first layer of dsn, give a warning and (unless quiet = TRUE) print a message when there are multiple layers, or give an error if there are no layers in dsn. If dsn is a database connection, then layer can be a table name or a database identifier (see Id). It is also possible to omit layer and rather use the query argument.

...

ignored

query

SQL query to pass in directly

Details

Lazy means that the usual behaviour of reading the entirety of a data source into memory is avoided. Printing the output results in a preview query being run and displayed (the top few rows of data).

The output of lazysf() is a 'tbl_SFSQLConnection⁠that extends⁠tbl_dbi' and may be used with functions and workflows in the normal DBI way, see SFSQL() for the lazysf DBI support.

The kind of q uery that may be run will depend on the type of format, see the list on the GDAL vector drivers page. For some details see the GDALSQL vignette.

When dplyr is attached the lazy data frame can be used with the usual verbs verbs (filter, select, distinct, mutate, transmute, arrange, left_join, pull, collect etc.). To see the result as a SQL query rather than a data frame preview use dplyr::show_query().

To obtain an in memory data frame use an explict collect() or st_as_sf(). A call to collect() is triggered by st_as_sf() and will add the sf class to the output. A result may not contain a geometry column, and so cannot be convert to an sf data frame. Using collect() on its own returns an unclassed data.frame and may include a classed sfc geometry column.

As well as collect() it's also possible to use tibble::as_tibble() or as.data.frame() or pull() which all force computation and retrieve the result.

Value

a 'tbl_SFSQLConnection', extending 'tbl_lazy' (something that works with dplyr verbs, and only shows a preview until you commit the result via collect()) see Details

Examples

# online sources can work
geojson <- file.path("https://raw.githubusercontent.com/SymbolixAU",
                     "geojsonsf/master/inst/examples/geo_melbourne.geojson")

lazysf(geojson)


## normal file stuff
## (Geopackage is an actual database so with SELECT we must be explicit re geom-column)
f <- system.file("gpkg/nc.gpkg", package = "sf", mustWork = TRUE)
lazysf(f)
lazysf(f, query = "SELECT AREA, FIPS, geom FROM \"nc.gpkg\" WHERE AREA < 0.1")
lazysf(f, layer = "nc.gpkg") %>% dplyr::select(AREA, FIPS, geom) %>% dplyr::filter(AREA < 0.1)

## the famous ESRI Shapefile (not an actual database)
## so if we SELECT we must be ex
shp <- lazysf(system.file("shape/nc.shp", package = "sf", mustWork = TRUE))
library(dplyr)
shp %>%
 filter(NAME %LIKE% 'A%') %>%
 mutate(abc = 1.3) %>%
 select(abc, NAME, `_ogr_geometry_`) %>%
 arrange(desc(NAME))  #%>% show_query()

 ## a multi-layer file
 system.file("extdata/multi.gpkg", package = "lazysf", mustWork = TRUE)