/
snowflake_io_manager.py
199 lines (175 loc) · 6.74 KB
/
snowflake_io_manager.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
from contextlib import contextmanager
from datetime import datetime
from typing import Any, Mapping, Optional, Sequence, Tuple, Union
from dagster import (
ConfigurableIOManager,
InputContext,
MetadataValue,
OutputContext,
TableColumn,
TableSchema,
)
from pandas import (
DataFrame as PandasDataFrame,
read_sql,
)
from pyspark.sql import DataFrame as SparkDataFrame
from snowflake.connector.pandas_tools import pd_writer
from snowflake.sqlalchemy import URL
from sqlalchemy import create_engine
SNOWFLAKE_DATETIME_FORMAT = "%Y-%m-%d %H:%M:%S"
@contextmanager
def connect_snowflake(config, schema="public"):
url = URL(
account=config["account"],
user=config["user"],
password=config["password"],
database=config["database"],
warehouse=config["warehouse"],
schema=schema,
timezone="UTC",
)
conn = None
try:
conn = create_engine(url).connect()
yield conn
finally:
if conn:
conn.close()
class SnowflakeIOManager(ConfigurableIOManager):
"""This IOManager can handle outputs that are either Spark or Pandas DataFrames. In either case,
the data will be written to a Snowflake table specified by metadata on the relevant Out.
"""
account: str
user: str
password: str
database: str
warehouse: str
@property
def _config(self):
return self.dict()
def handle_output(self, context: OutputContext, obj: Union[PandasDataFrame, SparkDataFrame]):
schema, table = context.asset_key.path[-2], context.asset_key.path[-1]
time_window = context.asset_partitions_time_window if context.has_asset_partitions else None
with connect_snowflake(config=self._config, schema=schema) as con:
con.execute(self._get_cleanup_statement(table, schema, time_window))
if isinstance(obj, SparkDataFrame):
metadata = self._handle_spark_output(obj, schema, table)
elif isinstance(obj, PandasDataFrame):
metadata = self._handle_pandas_output(obj, schema, table)
elif obj is None: # dbt
config = dict(self._config)
config["schema"] = schema
with connect_snowflake(config=config) as con:
df = read_sql(f"SELECT * FROM {context.name} LIMIT 5", con=con)
num_rows = con.execute(f"SELECT COUNT(*) FROM {context.name}").fetchone()
metadata = {
"data_sample": MetadataValue.md(df.to_markdown()),
"rows": num_rows,
}
else:
raise Exception(
"SnowflakeIOManager only supports pandas DataFrames and spark DataFrames"
)
context.add_output_metadata(
dict(
query=self._get_select_statement(
table,
schema,
None,
time_window,
),
**metadata,
)
)
def _handle_pandas_output(
self, obj: PandasDataFrame, schema: str, table: str
) -> Mapping[str, Any]:
from snowflake import connector
connector.paramstyle = "pyformat"
with connect_snowflake(config=self._config, schema=schema) as con:
with_uppercase_cols = obj.rename(str.upper, copy=False, axis="columns")
with_uppercase_cols.to_sql(
table,
con=con,
if_exists="append",
index=False,
method=pd_writer,
)
return {
"rows": obj.shape[0],
"dataframe_columns": MetadataValue.table_schema(
TableSchema(
columns=[
TableColumn(name=name, type=str(dtype)) # type: ignore # (bad stubs)
for name, dtype in obj.dtypes.items()
]
)
),
}
def _handle_spark_output(
self, df: SparkDataFrame, schema: str, table: str
) -> Mapping[str, Any]:
options = {
"sfURL": f"{self._config['account']}.snowflakecomputing.com",
"sfUser": self._config["user"],
"sfPassword": self._config["password"],
"sfDatabase": self._config["database"],
"sfSchema": schema,
"sfWarehouse": self._config["warehouse"],
"dbtable": table,
}
df.write.format("net.snowflake.spark.snowflake").options(**options).mode("append").save()
return {
"dataframe_columns": MetadataValue.table_schema(
TableSchema(
columns=[
TableColumn(name=field.name, type=field.dataType.typeName())
for field in df.schema.fields
]
)
)
}
def _get_cleanup_statement(
self, table: str, schema: str, time_window: Optional[Tuple[datetime, datetime]]
) -> str:
"""Returns a SQL statement that deletes data in the given table to make way for the output data
being written.
"""
if time_window:
return f"DELETE FROM {schema}.{table} {self._time_window_where_clause(time_window)}"
else:
return f"DELETE FROM {schema}.{table}"
def load_input(self, context: InputContext) -> PandasDataFrame:
asset_key = context.asset_key
schema, table = asset_key.path[-2], asset_key.path[-1]
with connect_snowflake(config=self._config) as con:
result = read_sql(
sql=self._get_select_statement(
table,
schema,
(context.definition_metadata or {}).get("columns"),
context.asset_partitions_time_window if context.has_asset_partitions else None,
),
con=con,
)
result.columns = map(str.lower, result.columns) # type: ignore # (bad stubs)
return result
def _get_select_statement(
self,
table: str,
schema: str,
columns: Optional[Sequence[str]],
time_window: Optional[Tuple[datetime, datetime]],
):
col_str = ", ".join(columns) if columns else "*"
if time_window:
return (
f"""SELECT {col_str} FROM {self._config["database"]}.{schema}.{table}\n"""
+ self._time_window_where_clause(time_window)
)
else:
return f"""SELECT {col_str} FROM {schema}.{table}"""
def _time_window_where_clause(self, time_window: Tuple[datetime, datetime]) -> str:
start_dt, end_dt = time_window
return f"""WHERE TO_TIMESTAMP(time::INT) BETWEEN '{start_dt.strftime(SNOWFLAKE_DATETIME_FORMAT)}' AND '{end_dt.strftime(SNOWFLAKE_DATETIME_FORMAT)}'"""