Merge pull request #171 from datafold/separated

Now separating the columns with a '|' before concatting them

Author: erezsh

data_diff/sql.py

@@ -6,6 +6,8 @@
 
 from runtype import dataclass
 
+from .utils import join_iter
+
 from .databases.database_types import AbstractDatabase, DbPath, DbKey, DbTime, ArithUUID
 
 
@@ -15,6 +17,8 @@ class Sql:
 
 SqlOrStr = Union[Sql, str]
 
+CONCAT_SEP = "|"
+
 
 @dataclass
 class Compiler:
@@ -122,7 +126,8 @@ class Checksum(Sql):
     def compile(self, c: Compiler):
         if len(self.exprs) > 1:
             compiled_exprs = [f"coalesce({c.compile(expr)}, '<null>')" for expr in self.exprs]
-            expr = c.database.concat(compiled_exprs)
+            separated = list(join_iter(f"'|'", compiled_exprs))
+            expr = c.database.concat(separated)
         else:
             # No need to coalesce - safe to assume that key cannot be null
             (expr,) = self.exprs

data_diff/utils.py

@@ -51,3 +51,11 @@ def number_to_human(n):
     )
 
     return "{:.0f}{}".format(n / 10 ** (3 * millidx), millnames[millidx])
+
+
+def join_iter(joiner: Any, iterable: iter) -> iter:
+    it = iter(iterable)
+    yield next(it)
+    for i in it:
+        yield joiner
+        yield i

tests/test_diff_tables.py

@@ -254,7 +254,7 @@ def test_get_values(self):
         table = self.table.with_schema()
 
         self.assertEqual(1, table.count())
-        concatted = str(id_) + time
+        concatted = str(id_) + "|" + time
         self.assertEqual(str_to_checksum(concatted), table.count_and_checksum()[1])
 
     def test_diff_small_tables(self):