ARROW-7045: [R] Preserve factor in Parquet roundtrip

The ability to preserve categorical values was introduced in #5077 as the convention of storing a special `ARROW:schema` key in the metadata. To invoke this, we need to call `ArrowWriterProperties::store_schema()`.

The R binding is already ready for this, but calls `store_schema()` only conditionally and uses `parquet___default_arrow_writer_properties()` by default. Though I don't see the motivation to implement as such in #5451, considering [the Python binding always calls `store_schema()`](https://github.com/apache/arrow/blob/dbe708c7527a4aa6b63df7722cd57db4e0bd2dc7/python/pyarrow/_parquet.pyx#L1269), I guess the R code can do the same.

Closes #6135 from yutannihilation/ARROW-7045_preserve_factor_in_parquet and squashes the following commits:

9227e7e <Hiroaki Yutani> Fix test
4d8bb46 <Hiroaki Yutani> Remove default_arrow_writer_properties()
dfd08cb <Hiroaki Yutani> Add failing tests

Authored-by: Hiroaki Yutani <[email protected]>
Signed-off-by: Neal Richardson <[email protected]>

Author: yutannihilation

r/R/arrowExports.R

@@ -205,16 +205,12 @@ ParquetArrowWriterPropertiesBuilder <- R6Class("ParquetArrowWriterPropertiesBuil
 ParquetArrowWriterProperties <- R6Class("ParquetArrowWriterProperties", inherit = Object)
 
 ParquetArrowWriterProperties$create <- function(use_deprecated_int96_timestamps = FALSE, coerce_timestamps = NULL, allow_truncated_timestamps = FALSE) {
-  if (!use_deprecated_int96_timestamps && is.null(coerce_timestamps) && !allow_truncated_timestamps) {
-    shared_ptr(ParquetArrowWriterProperties, parquet___default_arrow_writer_properties())
-  } else {
-    builder <- shared_ptr(ParquetArrowWriterPropertiesBuilder, parquet___ArrowWriterProperties___Builder__create())
-    builder$store_schema()
-    builder$set_int96_support(use_deprecated_int96_timestamps)
-    builder$set_coerce_timestamps(coerce_timestamps)
-    builder$set_allow_truncated_timestamps(allow_truncated_timestamps)
-    shared_ptr(ParquetArrowWriterProperties, parquet___ArrowWriterProperties___Builder__build(builder))
-  }
+  builder <- shared_ptr(ParquetArrowWriterPropertiesBuilder, parquet___ArrowWriterProperties___Builder__create())
+  builder$store_schema()
+  builder$set_int96_support(use_deprecated_int96_timestamps)
+  builder$set_coerce_timestamps(coerce_timestamps)
+  builder$set_allow_truncated_timestamps(allow_truncated_timestamps)
+  shared_ptr(ParquetArrowWriterProperties, parquet___ArrowWriterProperties___Builder__build(builder))
 }
 
 valid_parquet_version <- c(

r/R/parquet.R

@@ -82,12 +82,6 @@ std::shared_ptr<arrow::Table> parquet___arrow___FileReader__ReadTable2(
   return table;
 }
 
-// [[arrow::export]]
-std::shared_ptr<parquet::ArrowWriterProperties>
-parquet___default_arrow_writer_properties() {
-  return parquet::default_arrow_writer_properties();
-}
-
 // [[arrow::export]]
 std::shared_ptr<parquet::ArrowWriterProperties::Builder>
 parquet___ArrowWriterProperties___Builder__create() {

r/src/arrowExports.cpp

@@ -106,3 +106,17 @@ test_that("write_parquet() defaults to snappy compression", {
   write_parquet(mtcars, tmp2, compression = "snappy")
   expect_equal(file.size(tmp1), file.size(tmp2))
 })
+
+test_that("Factors are preserved when writing/reading from Parquet", {
+  fct <- factor(c("a", "b"), levels = c("c", "a", "b"))
+  ord <- factor(c("a", "b"), levels = c("c", "a", "b"), ordered = TRUE)
+  chr <- c("a", "b")
+  df <- tibble::tibble(fct = fct, ord = ord, chr = chr)
+
+  pq_tmp_file <- tempfile()
+  on.exit(unlink(pq_tmp_file))
+
+  write_parquet(df, pq_tmp_file)
+  df_read <- read_parquet(pq_tmp_file)
+  expect_identical(df, df_read)
+})