From fe64f4248587efbcf386f05a8be5b921a72b9567 Mon Sep 17 00:00:00 2001
From: Pradeep Krishnamurthy <pradkrish84@gmail.com>
Date: Tue, 25 May 2021 21:12:06 +0200
Subject: [PATCH] fread parse error (#2708)

At `src/core/read/fread/fread_thread_context.cc:210,` If `j<ncols` and if `*tch` reaches the end of line with the current quote rule, the while loop exits ending up in IOError.

With the code added in this PR, one way to fix this is to decrement tch`(tch--)` to the previous `sep` and try parsing again with the next quote rule. If the number of columns in the current row is only one, decrementing tch will never hit the sep in the current row and end up running into the previous row, hence the `if` logic inside the `while` loop to prevent this.

Closes #2680
---
 src/core/read/fread/fread_thread_context.cc | 15 +++++++++++++++
 tests/fread/test-fread-issues.py            |  5 +++++
 tests/fread/test-fread-large.py             |  4 +++-
 3 files changed, 23 insertions(+), 1 deletion(-)

diff --git a/src/core/read/fread/fread_thread_context.cc b/src/core/read/fread/fread_thread_context.cc
index f08dfe4cba..88c2171ed1 100644
--- a/src/core/read/fread/fread_thread_context.cc
+++ b/src/core/read/fread/fread_thread_context.cc
@@ -192,6 +192,21 @@ void FreadThreadContext::read_chunk(
         }
         parse_ctx_.target ++;
         j++;
+        if (tch < parse_ctx_.eof && j < ncols && *tch=='\n' && sep!=' ') {
+          const char* prev_tch = tch;
+          while (*tch!=sep) {
+            tch--;
+            if (*tch=='\n') {
+              tch = prev_tch;
+              break;
+            }
+          }
+          if (*tch==sep) {
+            tch++;
+            ++ptype_iter;
+            continue;
+          }
+        }
         if (tch < parse_ctx_.eof && *tch==sep) { tch++; continue; }
         if (fill && (tch == parse_ctx_.eof || *tch=='\n' || *tch=='\r') && j <= ncols) {
           // All parsers have already stored NA to target; except for string
diff --git a/tests/fread/test-fread-issues.py b/tests/fread/test-fread-issues.py
index bcd7f7d63c..3b78c58dea 100644
--- a/tests/fread/test-fread-issues.py
+++ b/tests/fread/test-fread-issues.py
@@ -542,6 +542,11 @@ def test_issue2523():
     with pytest.raises(IOError):
         dt.fread("{\n  \"cells\": [\n    {\n\"import numpy \\n\",\n")
 
+# Note: Realised after adding this test that it's similar to test_issue1036
+def test_issue2680():
+    src = '1\tWild Hogs (2007)\tAdevnture\n' * 500 + '2\t"Great Performances" Cats (1998)\tMusical\n' * 500
+    DT = dt.fread(src, fill=True)
+    assert DT.to_tuples()[900] == (2, '"Great Performances" Cats (1998)', 'Musical')
 
 def test_issue934():
     DT = dt.fread("A,B,C\n1,2,3\n3,4,5\n0,0,\"moo\n\n")
diff --git a/tests/fread/test-fread-large.py b/tests/fread/test-fread-large.py
index 5e3c10454c..831ce9c205 100644
--- a/tests/fread/test-fread-large.py
+++ b/tests/fread/test-fread-large.py
@@ -198,7 +198,9 @@ def test_h2o3_smalldata(f):
             return
         if "test_pubdev3589" in f:
             params["sep"] = "\n"
-        if "single_quotes_mixed.csv" in f or "single_quotes_with_escaped_quotes.csv" in f:
+        if ("single_quotes_mixed.csv" in f or
+            "single_quotes_with_escaped_quotes.csv" in f or
+            "single_quotes_with_escaped_quotes_custom_escapechar.csv" in f):
             params["quotechar"] = "'"
         with warnings.catch_warnings():
             warnings.simplefilter("ignore")