csv: add scientific notation without dots

pola-rs · Dec 4, 2021 · 0ca6ff8 · 0ca6ff8
1 parent 33cd8d1
commit 0ca6ff8
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 1 deletion.
diff --git a/polars/polars-io/src/csv.rs b/polars/polars-io/src/csv.rs
@@ -1344,4 +1344,18 @@ A3,\"B4_\"\"with_embedded_double_quotes\"\"\",C4,4";
 
         Ok(())
     }
+
+    #[test]
+    fn test_scientific_floats() -> Result<()> {
+        let csv = r#"foo,bar
+10000001,1e-5
+10000002,.04
+"#;
+        let file = Cursor::new(csv);
+        let df = CsvReader::new(file).finish()?;
+        assert_eq!(df.shape(), (2, 2));
+        assert_eq!(df.dtypes(), &[DataType::Int64, DataType::Float64]);
+
+        Ok(())
+    }
 }
diff --git a/polars/polars-io/src/csv_core/utils.rs b/polars/polars-io/src/csv_core/utils.rs
@@ -71,7 +71,8 @@ pub fn get_reader_bytes<R: Read + MmapBytesReader>(reader: &mut R) -> Result<Rea
 }
 
 lazy_static! {
-    static ref FLOAT_RE: Regex = Regex::new(r"^\s*-?((\d*\.\d+)[eE]?[-\+]?\d*)|inf|NaN$").unwrap();
+    static ref FLOAT_RE: Regex =
+        Regex::new(r"^\s*-?((\d*\.\d+)[eE]?[-\+]?\d*)|inf|NaN|\d+[eE][-+]\d+$").unwrap();
     static ref INTEGER_RE: Regex = Regex::new(r"^\s*-?(\d+)$").unwrap();
     static ref BOOLEAN_RE: Regex = RegexBuilder::new(r"^\s*(true)$|^(false)$")
         .case_insensitive(true)