diff --git a/crates/nu-command/src/formats/from/csv.rs b/crates/nu-command/src/formats/from/csv.rs
index 74e4bcb16..671217ca5 100644
--- a/crates/nu-command/src/formats/from/csv.rs
+++ b/crates/nu-command/src/formats/from/csv.rs
@@ -21,7 +21,7 @@ impl Command for FromCsv {
             .named(
                 "separator",
                 SyntaxShape::String,
-                "a character to separate columns, defaults to ','",
+                "a character to separate columns (either single char or 4 byte unicode sequence), defaults to ','",
                 Some('s'),
             )
             .named(
@@ -134,11 +134,22 @@ fn from_csv(
 ) -> Result<PipelineData, ShellError> {
     let name = call.head;
 
-    let separator = call
-        .get_flag(engine_state, stack, "separator")?
-        .map(|v: Value| v.as_char())
-        .transpose()?
-        .unwrap_or(',');
+    let separator = match call.get_flag::<String>(engine_state, stack, "separator")? {
+        Some(sep) => {
+            if sep.len() == 1 {
+                sep.chars().next().unwrap_or(',')
+            } else if sep.len() == 4 {
+                let unicode_sep = u32::from_str_radix(&sep, 16);
+                char::from_u32(unicode_sep.unwrap_or(b'\x1f' as u32)).unwrap_or(',')
+            } else {
+                return Err(ShellError::NonUtf8Custom(
+                    "separator should be a single char or a 4-byte unicode".to_string(),
+                    call.span(),
+                ));
+            }
+        }
+        None => ',',
+    };
     let comment = call
         .get_flag(engine_state, stack, "comment")?
         .map(|v: Value| v.as_char())
diff --git a/crates/nu-command/tests/format_conversions/csv.rs b/crates/nu-command/tests/format_conversions/csv.rs
index 43d1fba35..d609366a2 100644
--- a/crates/nu-command/tests/format_conversions/csv.rs
+++ b/crates/nu-command/tests/format_conversions/csv.rs
@@ -341,7 +341,9 @@ fn from_csv_text_with_multiple_char_separator() {
             "#
         ));
 
-        assert!(actual.err.contains("single character separator"));
+        assert!(actual
+            .err
+            .contains("separator should be a single char or a 4-byte unicode"));
     })
 }
 
@@ -366,7 +368,7 @@ fn from_csv_text_with_wrong_type_separator() {
             "#
         ));
 
-        assert!(actual.err.contains("can't convert int to char"));
+        assert!(actual.err.contains("can't convert int to string"));
     })
 }
 
@@ -404,3 +406,57 @@ fn string_to_csv_error() {
 
     assert!(actual.err.contains("command doesn't support"))
 }
+
+#[test]
+fn parses_csv_with_unicode_sep() {
+    Playground::setup("filter_from_csv_unicode_sep_test_3", |dirs, sandbox| {
+        sandbox.with_files(vec![FileWithContentToBeTrimmed(
+            "los_tres_caballeros.txt",
+            r#"
+                first_name;last_name;rusty_luck
+                Andrés;Robalino;1
+                JT;Turner;1
+                Yehuda;Katz;1
+            "#,
+        )]);
+
+        let actual = nu!(
+            cwd: dirs.test(), pipeline(
+            r#"
+                open los_tres_caballeros.txt
+                | from csv --separator "003B"
+                | get rusty_luck
+                | length
+            "#
+        ));
+
+        assert_eq!(actual.out, "3");
+    })
+}
+
+#[test]
+fn parses_csv_with_unicode_x1f_sep() {
+    Playground::setup("filter_from_csv_unicode_sep_x1f_test_3", |dirs, sandbox| {
+        sandbox.with_files(vec![FileWithContentToBeTrimmed(
+            "los_tres_caballeros.txt",
+            r#"
+                first_namelast_namerusty_luck
+                AndrésRobalino1
+                JTTurner1
+                YehudaKatz1
+            "#,
+        )]);
+
+        let actual = nu!(
+            cwd: dirs.test(), pipeline(
+            r#"
+                open los_tres_caballeros.txt
+                | from csv --separator "001F"
+                | get rusty_luck
+                | length
+            "#
+        ));
+
+        assert_eq!(actual.out, "3");
+    })
+}