From 3130d51f90e04c0bfdf7f5877a18a6f5566ab803 Mon Sep 17 00:00:00 2001
From: oliverbock <Oliver.bock@displayr.com>
Date: Fri, 9 Apr 2021 22:35:02 +1000
Subject: [PATCH] Use old-style SPSS compression by default for .sav (SPSS)
 files, since almost nobody uses the inefficient uncompressed format.  The
 newer ZCompressed format (compress=TRUE) is even better, but is much less
 used and less supported by third-party tools. (#544)

Co-authored-by: numbers <numbers@dockerhost>
---
 NEWS.md          | 4 ++++
 R/haven.R        | 2 +-
 man/read_spss.Rd | 2 +-
 src/DfWriter.cpp | 2 ++
 4 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/NEWS.md b/NEWS.md
index a584b38f..017c816e 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -15,6 +15,10 @@
 * Updated to ReadStat 1.1.5. Most importantly this includes support for
   SAS binary compression.
 
+* `write_sav(..., compress=FALSE)` now uses SPSS Bytecode compression instead
+  of the rarely-used uncompressed mode.  `compress=TRUE` continues to use the new
+  (and not universally supported, but more compact) ZLIB format. (@oliverbock, #544)
+
 # haven 2.3.1
 
 * Add missing methods so `median()`, `quantile()` and `summary()` work
diff --git a/R/haven.R b/R/haven.R
index ec34e910..35ca60d8 100644
--- a/R/haven.R
+++ b/R/haven.R
@@ -217,7 +217,7 @@ read_por <- function(file, user_na = FALSE, col_select = NULL, skip = 0, n_max =
 #' @export
 #' @rdname read_spss
 #' @param compress If `TRUE`, will compress the file, resulting in a `.zsav`
-#'   file.
+#'   file.  Otherwise the `.sav` file will be bytecode compressed.
 write_sav <- function(data, path, compress = FALSE) {
   validate_sav(data)
   write_sav_(data, normalizePath(path, mustWork = FALSE), compress = compress)
diff --git a/man/read_spss.Rd b/man/read_spss.Rd
index f7bed388..abb1ce30 100644
--- a/man/read_spss.Rd
+++ b/man/read_spss.Rd
@@ -91,7 +91,7 @@ to enforce them.}
 \item{path}{Path to a file where the data will be written.}
 
 \item{compress}{If \code{TRUE}, will compress the file, resulting in a \code{.zsav}
-file.}
+file.  Otherwise the \code{.sav} file will be bytecode compressed.}
 }
 \value{
 A tibble, data frame variant with nice defaults.
diff --git a/src/DfWriter.cpp b/src/DfWriter.cpp
index 52cef9c6..64bdaa14 100644
--- a/src/DfWriter.cpp
+++ b/src/DfWriter.cpp
@@ -403,6 +403,8 @@ void write_sav_(cpp11::list data, cpp11::strings path, bool compress) {
   Writer writer(HAVEN_SAV, data, path);
   if (compress)
     writer.setCompression(READSTAT_COMPRESS_BINARY);
+  else
+    writer.setCompression(READSTAT_COMPRESS_ROWS);
   writer.write();
 }