def assert_results(result: DataFrame) -> None:
"""
Shared asserts for the different formats of CSV file, all of which contain the same data.
"""
# Assert
assert result.count() == 3
assert result.collect()[1][0] == "2"
assert result.collect()[1][1] == "bar"
assert result.collect()[1][2] == "bar2"
def insert_df_xy(df: DataFrame,
name: str,
ws: str = "memory",
spatial_reference: int = 3857) -> None:
"""Create ephemeral point feature class from given dataframe.
Note - It is assume that the first two data fields are the point x/y values.
:param df: A dataframe.
:param name: The name of the feature class.
:param ws: The feature class workspace.
:param spatial_reference: The feature class spatial reference.
"""
fields = _df_to_fields(df, 2)
rows = df.collect()
insert_rows_xy(rows, name, fields, ws, spatial_reference)
def naaccr_read_fwf(flat_file: DataFrame,
record_layout: DataFrame,
value_col: str = 'value',
exclude_pfx: str = 'reserved') -> DataFrame:
"""
@param flat_file: as from spark.read.text()
typically with .value
@param record_layout: as from http://datadictionary.naaccr.org/?c=7
with .start, .length, .xmlId
"""
fields = [
func.substring(flat_file[value_col], item.start,
item.length).alias(item.xmlId)
for item in record_layout.collect()
if not item.xmlId.startswith(exclude_pfx)
] # type: List[Union[Column, str]]
return flat_file.select(fields)
def insert_df_hex(df: DataFrame,
name: str,
size: float,
ws: str = "memory") -> None:
"""Create ephemeral polygon feature class from given dataframe.
Note - It is assume that the first field is the hex nume value.
:param df: A dataframe.
:param name: The name of the feature class.
:param size: The hex size in meters.
:param ws: The feature class workspace.
"""
layout = Layout(size)
fields = _df_to_fields(df, 1)
rows = df.collect()
with insert_cursor(name, fields, ws=ws, shape_format="") as cursor:
for nume, *tail in rows:
coords = Hex.from_nume(nume).to_coords(layout)
cursor.insertRow((coords, *tail))
def insert_df(df: DataFrame,
name: str,
ws: str = "memory",
spatial_reference: int = 3857,
shape_type: str = "POLYGON",
shape_format: str = "WKB") -> None:
"""Create an ephemeral feature class given a dataframe.
Note - it is assumed that the first data field is the shape field.
:param df: A dataframe.
:param name: The name of the feature class.
:param ws: The output workspace. Default="memory".
:param spatial_reference: The spatial reference id. Default=2857.
:param shape_type: The feature class shape type (POINT,POLYGON,POLYLINE,MULTIPOINT). Default="POLYGON".
:param shape_format: The shape format (WKB, WKT, ''). Default="WKB".
"""
fields = _df_to_fields(df, 1)
rows = df.collect()
insert_rows(rows, name, fields, ws, spatial_reference, shape_type,
shape_format)
def synthesize_data(self,
stats_nom: DataFrame,
record_layout: DataFrame,
qty: int = 100) -> pd.DataFrame:
spark = self.__spark
create_object(self.t_item_view, self.concepts_script, spark)
create_object(self.txform_view, self.txform_script, spark)
stats_nom.createOrReplaceTempView(self.agg_view)
entity = spark.createDataFrame([(ix, ) for ix in range(0, qty)],
['case_index'])
entity.createOrReplaceTempView(self.entity_view)
# simulated_entity.limit(5).toPandas()
for view in self.views:
create_object(view, self.script, spark)
spark.catalog.cacheTable(self.views[-1])
# ISSUE: SQL goes in .sql files
sim_records_nom = spark.sql('''
select data.case_index, data.xmlId, data.value
from simulated_naaccr_nom data
join record_layout rl on rl.xmlId = data.xmlId
join section on rl.section = section.section
order by case_index, rl.start
''').toPandas()
sim_records_nom = sim_records_nom.pivot(index='case_index',
columns='xmlId',
values='value')
for col in sim_records_nom.columns:
sim_records_nom[col] = sim_records_nom[col].astype('category')
col_start = {row.xmlId: row.start for row in record_layout.collect()}
sim_records_nom = sim_records_nom[sorted(
sim_records_nom.columns, key=lambda xid: col_start[xid])]
return sim_records_nom
