def test_create_select_table_range_datetime_index(empty_db, simple_df,
constants):
"""add a new table with explicit index, read it back with pandabase, check equality"""
simple_df.index = simple_df.date
simple_df = simple_df.drop('date', axis=1)
table = pb.to_sql(simple_df,
table_name='sample',
con=empty_db,
how='create_only')
# print(table.columns)
assert table.columns['date'].primary_key
assert pb.has_table(empty_db, 'sample')
loaded0 = pb.read_sql('sample',
con=empty_db,
lowest=simple_df.index[-1],
highest=simple_df.index[0])
print(loaded0)
assert len(loaded0) == 0
loaded = pb.read_sql('sample',
con=empty_db,
lowest=simple_df.index[0],
highest=simple_df.index[-1])
assert pb.companda(loaded, simple_df, ignore_all_nan_columns=True)
loaded = pb.read_sql('sample', con=empty_db, highest=simple_df.index[-1])
assert pb.companda(loaded, simple_df, ignore_all_nan_columns=True)
loaded = pb.read_sql('sample', con=empty_db, lowest=simple_df.index[0])
assert pb.companda(loaded, simple_df, ignore_all_nan_columns=True)
def test_create_table_no_index_load_pandas(empty_db, minimal_df, as_type, how):
"""add a new minimal table to db, read with Pandas"""
if as_type:
minimal_df['integer'] = minimal_df['integer'].astype('Int64')
table = pb.to_sql(
minimal_df,
table_name='sample',
con=empty_db,
how=how,
auto_index=True,
)
# print(table.columns)
assert table.columns[PANDABASE_DEFAULT_INDEX].primary_key
assert pb.has_table(empty_db, 'sample')
loaded = pd.read_sql_table('sample',
con=empty_db,
index_col=PANDABASE_DEFAULT_INDEX)
# pandas doesn't know about default index
loaded.index.name = None
# pandas doesn't know stored as UTC w/o timezone info
loaded.date = pd.to_datetime(loaded.date, utc=True)
assert pb.companda(loaded, minimal_df, ignore_index=True)
def test_create_select_table_index(session_db, simple_df, constants):
"""add a new table with explicit index, read it back with pandabase, check equality"""
table = pb.to_sql(simple_df,
table_name='sample',
con=session_db,
how='create_only')
# print(table.columns)
assert table.columns[constants.SAMPLE_INDEX_NAME].primary_key
assert pb.has_table(session_db, 'sample')
loaded = pb.read_sql('sample', con=session_db)
assert pb.companda(loaded, simple_df, ignore_all_nan_columns=True)
def test_create_read_table_no_index(empty_db, minimal_df):
"""add a new minimal table & read it back with pandabase"""
table = pb.to_sql(minimal_df,
table_name='sample',
con=empty_db,
how='create_only',
auto_index=True,
)
# print(table.columns)
assert table.columns[PANDABASE_DEFAULT_INDEX].primary_key
loaded = pb.read_sql('sample', con=empty_db)
assert pb.has_table(empty_db, 'sample')
assert pb.companda(loaded, minimal_df, ignore_index=True)
def test_append_auto_index(empty_db, minimal_df):
"""add a new minimal table; add it again"""
pb.to_sql(minimal_df,
table_name='sample',
con=empty_db,
auto_index=True,
how='create_only')
table2 = pb.to_sql(minimal_df,
table_name='sample',
con=empty_db,
auto_index=True,
how='append')
assert table2.columns[PANDABASE_DEFAULT_INDEX].primary_key
loaded = pb.read_sql('sample', con=empty_db)
assert pb.has_table(empty_db, 'sample')
double_df = pd.concat([minimal_df, minimal_df], ignore_index=True)
assert pb.companda(loaded, double_df, ignore_index=True)
assert len(loaded) == len(minimal_df) * 2
def test_create_read_table_with_different_index(session_db, simple_df, table_name, index_col_name):
"""create new tables in empty db, using different col types as index, read with pandabase"""
orig_df = simple_df.copy()
orig_df.index = orig_df[index_col_name]
print(orig_df[index_col_name])
print(orig_df.index)
orig_df = orig_df.drop(index_col_name, axis=1)
table = pb.to_sql(orig_df,
table_name=table_name,
con=session_db,
how='create_only')
assert table.columns[index_col_name].primary_key
assert pb.has_table(session_db, table_name)
loaded = pb.read_sql(table_name, con=session_db)
c = pb.companda(loaded, orig_df, ignore_all_nan_columns=True)
if not c:
raise ValueError(c.message)
def test_create_table_with_different_index_pandas(session_db, simple_df,
table_name, index_col_name):
"""create new tables in empty db, using different col types as index, read with Pandas"""
df = simple_df.copy()
df.index = df[index_col_name]
df = df.drop(index_col_name, axis=1)
table = pb.to_sql(df,
table_name=table_name,
con=session_db,
how='create_only')
assert table.columns[index_col_name].primary_key
assert pb.has_table(session_db, table_name)
# read with PANDAS
loaded = pd.read_sql_table(table_name,
con=session_db,
index_col=index_col_name)
# make an integer index, since pd.read_sql_table doesn't know to do this
new_index = loaded.index.name
loaded[new_index] = loaded.index
if isinstance(loaded[new_index].iloc[0], str):
print('converting')
loaded[new_index] = loaded[new_index].apply(lambda x: float(x))
loaded.index = loaded[new_index]
loaded = loaded.drop(new_index, axis=1)
# pandas doesn't know about UTC
if 'date' in loaded.columns:
print('converting date to UTC')
loaded.date = pd.to_datetime(loaded.date, utc=True)
else:
print('making new UTC index (Fake!)')
loaded.index = df.index
c = pb.companda(loaded, df, ignore_all_nan_columns=True)
if not c:
raise ValueError(c.message)
