add skip_blank_lines option

csv-core/src/reader.rs

@@ -115,6 +115,8 @@ pub struct Reader {
     /// If enabled (the default), then quotes are respected. When disabled,
     /// quotes are not treated specially.
     quoting: bool,
+    /// If enabled (the default) blank lines are ignored.
+    skip_blank_lines: bool,
     /// Whether to use the NFA for parsing.
     ///
     /// Generally this is for debugging. There's otherwise no good reason
@@ -141,6 +143,7 @@ impl Default for Reader {
             double_quote: true,
             comment: None,
             quoting: true,
+            skip_blank_lines: true,
             use_nfa: false,
             line: 1,
             has_read: false,
@@ -227,6 +230,16 @@ impl ReaderBuilder {
         self
     }
 
+    /// Enable or disable skipping of blank lines
+    ///
+    /// This is enabled by default, but it may be disabled. When enabled,
+    /// blank lines are ignored. If present, the finally trailing blank line
+    /// will be ignored.
+    pub fn skip_blank_lines(&mut self, yes: bool) -> &mut ReaderBuilder {
+        self.rdr.skip_blank_lines = yes;
+        self
+    }
+
     /// The comment character to use when parsing CSV.
     ///
     /// If the start of a record begins with the byte given here, then that
@@ -992,7 +1005,11 @@ impl Reader {
             End => (End, NfaInputAction::Epsilon),
             StartRecord => {
                 if self.term.equals(c) {
-                    (StartRecord, NfaInputAction::Discard)
+                    if self.skip_blank_lines {
+                        (StartRecord, NfaInputAction::Discard)
+                    } else {
+                        (CRLF, NfaInputAction::Discard)
+                    }
                 } else if self.comment == Some(c) {
                     (InComment, NfaInputAction::Discard)
                 } else {
@@ -1726,6 +1743,15 @@ mod tests {
         }
     );
 
+    fn enable_blank(builder: &mut ReaderBuilder) -> &mut ReaderBuilder {
+        builder.skip_blank_lines(false)
+    }
+    parses_to!(blank_lines_one_row_one_field, "a", csv![["a"]], enable_blank);
+    parses_to!(blank_lines_one_row_one_field_lf, "a\n", csv![["a"]], enable_blank);
+    parses_to!(blank_lines_one_row_one_field_lf_lf, "a\n\n", csv![["a"], [""]], enable_blank);
+    parses_to!(blank_lines_one_row_lf_one_field_lf, "\na\n", csv![[""], ["a"]], enable_blank);
+    parses_to!(blank_lines_crlf, "\r\na\r\n", csv![[""], ["a"]], enable_blank);
+
     macro_rules! assert_read {
         (
             $rdr:expr, $input:expr, $output:expr,

src/reader.rs

@@ -362,6 +362,41 @@ impl ReaderBuilder {
         self
     }
 
+    /// Whether blank lines are ignored.
+    ///
+    /// By default blank lines are ignored.
+    ///
+    /// When disabled 
+    ///
+    /// # Example
+    ///
+    /// ```
+    /// use std::error::Error;
+    /// use csv::ReaderBuilder;
+    ///
+    /// # fn main() { example().unwrap(); }
+    /// fn example() -> Result<(), Box<dyn Error>> {
+    ///     let data = "\na,b";
+    ///     let mut rdr = ReaderBuilder::new()
+    ///         .skip_blank_lines(false)
+    ///         .flexible(true)
+    ///         .has_headers(false)
+    ///         .from_reader(data.as_bytes());
+    ///
+    ///     if let Some(result) = rdr.records().next() {
+    ///         let record = result?;
+    ///         assert_eq!(record, vec![""]);
+    ///         Ok(())
+    ///     } else {
+    ///         Err(From::from("expected at least one record but got none"))
+    ///     }
+    /// }
+    /// ```
+    pub fn skip_blank_lines(&mut self, yes: bool) -> &mut ReaderBuilder {
+        self.builder.skip_blank_lines(yes);
+        self
+    }
+
     /// The record terminator to use when parsing CSV.
     ///
     /// A record terminator can be any single byte. The default is a special
@@ -2508,6 +2543,32 @@ mod tests {
         assert_eq!("baz", &headers[2]);
     }
 
+    #[test]
+    fn read_record_blank_lines() {
+        let data = b("foo,bar,baz\n\na,b,c\nd,e,f");
+        let mut rdr = ReaderBuilder::new()
+            .flexible(true).skip_blank_lines(false).has_headers(false).from_reader(data);
+        let mut rec = StringRecord::new();
+
+        assert!(rdr.read_record(&mut rec).unwrap());
+        assert_eq!(3, rec.len());
+        assert_eq!("foo", &rec[0]);
+
+        assert!(rdr.read_record(&mut rec).unwrap());
+        assert_eq!(1, rec.len());
+        assert_eq!("", &rec[0]);
+
+        assert!(rdr.read_record(&mut rec).unwrap());
+        assert_eq!(3, rec.len());
+        assert_eq!("a", &rec[0]);
+
+        assert!(rdr.read_record(&mut rec).unwrap());
+        assert_eq!(3, rec.len());
+        assert_eq!("d", &rec[0]);
+
+        assert!(!rdr.read_record(&mut rec).unwrap());
+    }
+
     #[test]
     fn seek() {
         let data = b("foo,bar,baz\na,b,c\nd,e,f\ng,h,i");