Do binomial check in label_encode functions only when adding new labe…

…ls (#1892)

For binomial problem number of labels should be less or equal to two. This check should only be performed when a new label is encountered in the target column. When running label encoding in parallel, we now do the check making sure the label was not added from another thread.

Closes #1891

CHANGELOG.md

@@ -209,7 +209,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/).
 - Thanks to everyone who helped make `datatable` more stable by discovering
   and reporting bugs that were fixed in this release:
 
-  - [Arno Candel][] (#1619, #1730, #1738, #1800, #1803, #1846, #1857),
+  - [Arno Candel][] (#1619, #1730, #1738, #1800, #1803, #1846, #1857, #1891),
   - [Antorsae][] (#1639),
   - [Olivier][] (#1872),
   - [Hawk Berry][] (#1834),

c/models/label_encode.h

@@ -124,12 +124,11 @@ void label_encode_fw(const Column* col, dtptr& dt_labels, dtptr& dt_encoded) {
         outdata[irow] = labels_map[v];
       } else {
         lock.exclusive_start();
-        if (stype_to == SType::BOOL && labels_map.size() == 2) {
-          throw ValueError() << "Target column for binomial problem cannot "
-                                "contain more than two labels";
-        }
-
         if (labels_map.count(v) == 0) {
+          if (stype_to == SType::BOOL && labels_map.size() == 2) {
+            throw ValueError() << "Target column for binomial problem cannot "
+                                  "contain more than two labels";
+          }
           size_t nlabels = labels_map.size();
           labels_map[v] = static_cast<T_to>(nlabels);
           outdata[irow] = labels_map[v];
@@ -192,12 +191,11 @@ void label_encode_str(const Column* col, dtptr& dt_labels, dtptr& dt_encoded) {
         outdata[irow] = labels_map[v];
       } else {
         lock.exclusive_start();
-        if (stype_to == SType::BOOL && labels_map.size() == 2) {
-          throw ValueError() << "Target column for binomial problem cannot "
-                                "contain more than two labels";
-        }
-
         if (labels_map.count(v) == 0) {
+          if (stype_to == SType::BOOL && labels_map.size() == 2) {
+            throw ValueError() << "Target column for binomial problem cannot "
+                                  "contain more than two labels";
+          }
           size_t nlabels = labels_map.size();
           labels_map[v] = static_cast<T_to>(nlabels);
           outdata[irow] = labels_map[v];