Improve detectDates. closes #288

DESCRIPTION

@@ -1,7 +1,7 @@
 Type: Package
 Package: openxlsx
 Title: Read, Write and Edit xlsx Files
-Version: 4.2.5
+Version: 4.2.5.1
 Date: 2021-12-13
 Authors@R: 
     c(person(given = "Philipp",

NEWS.md

@@ -1,3 +1,9 @@
+# openxlsx (development version)
+
+## Improvements
+
+* Improve detectDates ([#288](https://github.com/ycphs/openxlsx/issues/288))
+
 # openxlsx 4.2.5
 
 ## Fixes

src/write_file.cpp

@@ -1,5 +1,6 @@
 
 #include "openxlsx.h"
+#include <regex>
 
 
 //' @import Rcpp
@@ -197,7 +198,8 @@ SEXP buildMatrixMixed(CharacterVector v,
   for(int i = 0;i < k; i++)
     m(rowInd[i], colInd[i]) = v[i];
 
-
+  // convert numerics to dates
+  Rcpp::Function cTD("convertToDate");
 
   // this will be the return data.frame
   List dfList(nCols); 
@@ -225,16 +227,20 @@ SEXP buildMatrixMixed(CharacterVector v,
         if(!notNAElements[ri]){
           datetmp[ri] = NA_REAL; //IF TRUE, TRUE else FALSE
         }else{
-          // dt_str = as<std::string>(m(ri,i));
           dt_str = m(ri,i);
+
           try{
-            datetmp[ri] = Rcpp::Date(atoi(dt_str.substr(5,2).c_str()), atoi(dt_str.substr(8,2).c_str()), atoi(dt_str.substr(0,4).c_str()) );
-          }catch(...) {
+            // Some values are incorrectly detected as dates. This regex determines if they are numerics.
+            // If so, they are converted to Dates. 
+            if (std::regex_match( dt_str, std::regex( ( "((\\+|-)?[[:digit:]]+)(\\.(([[:digit:]]+)?))?" ) ) )) {
+              datetmp[ri] = as<Rcpp::Date>(cTD(dt_str));
+            } else {
+              datetmp[ri] = Rcpp::Date(atoi(dt_str.substr(5,2).c_str()), atoi(dt_str.substr(8,2).c_str()), atoi(dt_str.substr(0,4).c_str()) );
+            }
+          } catch(...) {
             Rcpp::Rcerr << "Error reading date:\n" << dt_str << "\nrow: " << ri+1 << "\ncol: " << i+1 << "\n";
             throw;
           }
-          //datetmp[ri] = Date(atoi(m(ri,i)) - originAdj);
-          //datetmp[ri] = Date(as<std::string>(m(ri,i)));
         }
       }
 

tests/testthat/test-date_time_conversion.R

@@ -1,9 +1,7 @@
 
-
 context("Date/Time Conversions")
 
 
-
 test_that("convert to date", {
   dates <- as.Date("2015-02-07") + -10:10
   origin <- 25569L
@@ -18,7 +16,6 @@ test_that("convert to date", {
 })
 
 
-
 test_that("convert to datetime", {
   x <- 43037 + 2 / 1440
   expect_equal(object = convertToDateTime(x, tx = Sys.timezone()), expected = as.POSIXct("2017-10-29 00:02:00", tz = Sys.timezone()))
@@ -33,3 +30,19 @@ test_that("convert to datetime", {
   x_datetime <- convertToDateTime(x, tx = "UTC")
   attr(x_datetime, "tzone") <- "UTC"
 })
+
+
+test_that("read.xlsx detectDates", {
+
+  xlsxFile <- system.file("extdata", "gh_issue_288.xlsx", package = "openxlsx")
+
+  ref_dat <- data.frame(Date = c(as.Date(c("2021-10-20", "2021-11-03"))))
+  ref_num <- data.frame(Date = c(44489.4, 44503.0))
+
+  tst_dat <- read.xlsx(xlsxFile, detectDates = TRUE)
+  tst_num <- read.xlsx(xlsxFile, detectDates = FALSE)
+
+  expect_equal(ref_dat, tst_dat)
+  expect_equal(ref_num, tst_num)
+
+})