Initial commit

.Rbuildignore

@@ -0,0 +1,10 @@
+^_targets\.R$
+^\.Rprofile$
+^report$
+^report/_bookdown\.yml$
+^report/_output\.yml$
+^report/index\.Rmd$
+^report/chapter1\.Rmd$
+^report/chapter2\.Rmd$
+^.*\.Rproj$
+^\.Rproj\.user$

.Rprofile

@@ -0,0 +1,36 @@
+if (file.exists("~/.Rprofile")) source("~/.Rprofile")
+
+if(!requireNamespace("pacman", quietly = TRUE)) {
+  utils::install.packages("pacman")
+}
+
+pacman::p_load(
+  magrittr,
+  dplyr,
+  tidyr,
+  ggplot2,
+
+  duckdb,
+  arrow,
+  DBI,
+  odbc,
+
+  targets,
+  tarchetypes
+)
+
+library(conflicted)
+conflicted::conflict_prefer("filter", "dplyr", quiet = TRUE)
+conflicted::conflict_prefer("select", "dplyr", quiet = TRUE)
+conflicted::conflict_prefer("expand", "tidyr", quiet = TRUE)
+
+# make all fns available to {targets} (and interactively, via load_all)
+if (rlang::is_interactive()) {
+  devtools::load_all()
+  # delayedAssign("duckdb_con", connect_duckdb(), assign.env = globalenv())
+  # delayedAssign("pg_con", connect_pg(), assign.env = globalenv())
+} else {
+  invisible(
+    lapply(list.files("./R", full.names = TRUE), source, encoding = "UTF-8")
+  )
+}

.gitignore

@@ -0,0 +1,9 @@
+.Rproj.user
+.Rhistory
+.RData
+.Ruserdata
+_targets
+playground*
+.Rdata
+.httr-oauth
+.DS_Store

DESCRIPTION

@@ -0,0 +1,37 @@
+Package: backend_compare
+Title: A {targets} pipeline to play around with {arrow} and {duckdb}
+Version: 0.0.0.9000
+Authors@R: 
+    person("Eduardo", "Alfonso-Sierra", , "[email protected]", 
+           role = c("aut", "cre", "cph"),
+           comment = c(ORCID = "0000-0001-6430-5135"))
+Description: Not really a package. Just a bunch of simple data wrangling steps
+             to play around with {arrow} and {duckdb}, and compare it with 
+             the same thing in PostgreSQL
+License: MIT + file LICENSE
+Encoding: UTF-8
+Language: en
+Roxygen: list(markdown = TRUE)
+RoxygenNote: 7.1.2
+Imports: 
+    targets,
+    tarchetypes,
+    visNetwork,
+    duckdb,
+    arrow,
+    conflicted,
+    dplyr,
+    ggplot2,
+    magrittr,
+    pacman,
+    rlang,
+    tidyr,
+    utils,
+    DBI,
+    odbc,
+    fs,
+    bench,
+    efun,
+    glue,
+    here,
+    purrr

NAMESPACE

@@ -0,0 +1,2 @@
+# Generated by roxygen2: do not edit by hand
+

R/aggregate.R

@@ -0,0 +1,3 @@
+#backend
+#interface
+

R/connect.R

@@ -0,0 +1,44 @@
+
+connect_duckdb <- function(read_only = FALSE) {
+
+  print("CONNECTING DuckDB ...")
+
+  fs::dir_create(fs::path_dir("/backend/duckdb/backend.duckdb"))
+  DBI::dbConnect(
+    drv = duckdb::duckdb(), 
+    dbdir = "/backend/duckdb/backend.duckdb",
+    read_only = read_only
+  )
+}
+
+connect_pg <- function() {
+
+  DBI::dbConnect(
+    odbc::odbc(),
+    driver = "PostgreSQL Unicode(x64)",
+    database = "backend",
+    uid = "postgres",
+    pwd = "postgres",
+    host = NULL, #,"localhost",
+    port = 5432,
+    encoding = "WINDOWS-1252",
+    bigint = "numeric"
+  )
+}
+
+connect_arrow <- function(arrow_tbl) {
+
+  arrow::open_dataset(
+    sources = arrow_tbl,
+    format = "arrow"
+  )
+}
+
+connect_parquet <- function(parquet_tbl) {
+
+  arrow::open_dataset(
+    sources = parquet_tbl,
+    format = "parquet"
+  )
+}
+