Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix DataFrame.merge to work properly #2060

Merged
merged 4 commits into from
Feb 22, 2021
Merged

Fix DataFrame.merge to work properly #2060

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
c7a1ad7
Fix DataFrame.merge to work properly
itholic Feb 19, 2021
b260942
preserve the Spark column name
itholic Feb 19, 2021
3a051f9
Addressed comments
itholic Feb 21, 2021
16c664d
Retrigger the test
itholic Feb 22, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
43 changes: 39 additions & 4 deletions databricks/koalas/frame.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7260,10 +7260,37 @@ def to_list(os: Optional[Union[Any, List[Any], Tuple, List[Tuple]]]) -> List[Tup
if len(left_key_names) != len(right_key_names):
raise ValueError("len(left_keys) must equal len(right_keys)")

# We should distinguish the name to avoid ambiguous column name after merging.
right_prefix = "__right_"
right_key_names = [right_prefix + right_key_name for right_key_name in right_key_names]

how = validate_how(how)

def resolve(internal, side):
rename = lambda col: "__{}_{}".format(side, col)
internal = internal.resolved_copy
sdf = internal.spark_frame
sdf = internal.spark_frame.select(
[
scol_for(sdf, col).alias(rename(col))
for col in sdf.columns
if col not in HIDDEN_COLUMNS
]
+ list(HIDDEN_COLUMNS)
)
return internal.copy(
spark_frame=sdf,
index_spark_columns=[
scol_for(sdf, rename(col)) for col in internal.index_spark_column_names
],
data_spark_columns=[
scol_for(sdf, rename(col)) for col in internal.data_spark_column_names
],
preserve_dtypes=True,
)

left_internal = self._internal.resolved_copy
right_internal = right._internal.resolved_copy
right_internal = resolve(right._internal, "right")

left_table = left_internal.spark_frame.alias("left_table")
right_table = right_internal.spark_frame.alias("right_table")
Expand Down Expand Up @@ -7301,9 +7328,13 @@ def to_list(os: Optional[Union[Any, List[Any], Tuple, List[Tuple]]]) -> List[Tup
scol = left_scol_for(label)
if label in duplicate_columns:
spark_column_name = left_internal.spark_column_name_for(label)
if spark_column_name in left_key_names and spark_column_name in right_key_names:
if (
spark_column_name in left_key_names
and (right_prefix + spark_column_name) in right_key_names
):
right_scol = right_scol_for(label)
if how == "right":
col = right_prefix + col
scol = right_scol
itholic marked this conversation as resolved.
Show resolved Hide resolved
elif how == "full":
scol = F.when(scol.isNotNull(), scol).otherwise(right_scol).alias(col)
Expand All @@ -7321,10 +7352,14 @@ def to_list(os: Optional[Union[Any, List[Any], Tuple, List[Tuple]]]) -> List[Tup
scol = right_scol_for(label)
if label in duplicate_columns:
spark_column_name = left_internal.spark_column_name_for(label)
if spark_column_name in left_key_names and spark_column_name in right_key_names:
if (
spark_column_name in left_key_names
and (right_prefix + spark_column_name) in right_key_names
):
continue
else:
col = col + right_suffix
# remove `right_prefix` here.
col = (col + right_suffix)[len(right_prefix) :]
scol = scol.alias(col)
label = tuple([str(label[0]) + right_suffix] + list(label[1:]))
itholic marked this conversation as resolved.
Show resolved Hide resolved
exprs.append(scol)
Expand Down
48 changes: 48 additions & 0 deletions databricks/koalas/tests/test_dataframe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2030,6 +2030,54 @@ def check(op, right_kdf=right_kdf, right_pdf=right_pdf):
)
)

def test_merge_same_anchor(self):
pdf = pd.DataFrame(
{
"lkey": ["foo", "bar", "baz", "foo", "bar", "l"],
"rkey": ["baz", "foo", "bar", "baz", "foo", "r"],
"value": [1, 1, 3, 5, 6, 7],
"x": list("abcdef"),
"y": list("efghij"),
},
columns=["lkey", "rkey", "value", "x", "y"],
)
kdf = ks.from_pandas(pdf)

left_pdf = pdf[["lkey", "value", "x"]]
right_pdf = pdf[["rkey", "value", "y"]]
left_kdf = kdf[["lkey", "value", "x"]]
right_kdf = kdf[["rkey", "value", "y"]]

def check(op, right_kdf=right_kdf, right_pdf=right_pdf):
k_res = op(left_kdf, right_kdf)
k_res = k_res.to_pandas()
k_res = k_res.sort_values(by=list(k_res.columns))
k_res = k_res.reset_index(drop=True)
p_res = op(left_pdf, right_pdf)
p_res = p_res.sort_values(by=list(p_res.columns))
p_res = p_res.reset_index(drop=True)
self.assert_eq(k_res, p_res)

check(lambda left, right: left.merge(right))
check(lambda left, right: left.merge(right, on="value"))
check(lambda left, right: left.merge(right, left_on="lkey", right_on="rkey"))
check(lambda left, right: left.set_index("lkey").merge(right.set_index("rkey")))
check(
lambda left, right: left.set_index("lkey").merge(
right, left_index=True, right_on="rkey"
)
)
check(
lambda left, right: left.merge(
right.set_index("rkey"), left_on="lkey", right_index=True
)
)
check(
lambda left, right: left.set_index("lkey").merge(
right.set_index("rkey"), left_index=True, right_index=True
)
)

def test_merge_retains_indices(self):
left_pdf = pd.DataFrame({"A": [0, 1]})
right_pdf = pd.DataFrame({"B": [1, 2]}, index=[1, 2])
Expand Down