Parse numbers with leading zeros as character.

Fixes #266

NEWS.md

@@ -1,5 +1,8 @@
 # readr 0.1.1.9000
 
+* Numbers with leading zeros now default to being parsed as text, rather than
+  as integers/doubles. Override with `col_integer()` or `col_double()` (#266).
+
 * `read_fwf()` is now much more careful with new lines. The last column 
   can now be ragged: the width of the last field is silently extended until
   it hits the next line break (#146). If a line is too short, you'll get

src/CollectorGuess.cpp

@@ -38,6 +38,8 @@ bool isLogical(const std::string& x, LocaleInfo* pLocale) {
 bool isInteger(const std::string& x, LocaleInfo* pLocale) {
   int res = 0;
   std::string::const_iterator begin = x.begin(), end = x.end();
+  if (x[0] == '0' && x.size() > 1)
+    return false;
 
   return parseInt(begin, end, res) && begin == end;
 }
@@ -55,6 +57,10 @@ bool canParseNumber(CharacterVector x, LocaleInfo* pLocale) {
     std::string xstr = std::string(x[i]);
     std::string::const_iterator begin = xstr.begin(), end = xstr.end();
 
+    // Leading zero not followed by decimal mark
+    if (xstr[0] == '0' && xstr.size() > 1 && xstr[1] != pLocale->decimalMark_)
+      return false;
+
     bool ok = parseNumber(pLocale->decimalMark_, pLocale->groupingMark_,
       begin, end, tmp);
     if (!ok)
@@ -69,6 +75,10 @@ bool canParseNumber(CharacterVector x, LocaleInfo* pLocale) {
 }
 
 bool isDouble(const std::string& x, LocaleInfo* pLocale) {
+  // Leading zero not followed by decimal mark
+  if (x[0] == '0' && x.size() > 1 && x[1] != pLocale->decimalMark_)
+    return false;
+
   double res = 0;
   std::string::const_iterator begin = x.begin(), end = x.end();
 

tests/testthat/test-parsing-numeric.R

@@ -20,6 +20,14 @@ test_that("leading/trailing ws ignored when parsing", {
 })
 
 
+# Leading zeros -----------------------------------------------------------
+
+test_that("leading zeros are not numbers", {
+  expect_equal(collector_guess("0"), "integer")
+  expect_equal(collector_guess("0."), "double")
+  expect_equal(collector_guess("0001"), "character")
+})
+
 # Flexible number parsing -------------------------------------------------
 es_MX <- locale("es", decimal_mark = ",")