def to_featureclass(geo,
location,
overwrite=True,
validate=False,
sanitize_columns=True,
has_m=True,
has_z=False):
"""
Exports the DataFrame to a Feature class.
=============== ====================================================
**Argument** **Description**
--------------- ----------------------------------------------------
location Required string. This is the output location for the
feature class. This should be the path and feature
class name.
--------------- ----------------------------------------------------
overwrite Optional Boolean. If overwrite is true, existing
data will be deleted and replaced with the spatial
dataframe.
--------------- ----------------------------------------------------
validate Optional Boolean. If True, the export will check if
all the geometry objects are correct upon export.
--------------- ----------------------------------------------------
sanitize_columns Optional Boolean. If True, column names will be
converted to string, invalid characters removed and
other checks will be performed. The default is True.
--------------- ----------------------------------------------------
ham_m Optional Boolean to indicate if data has linear
referencing (m) values. Default is False.
--------------- ----------------------------------------------------
has_z Optional Boolean to indicate if data has elevation
(z) values. Default is False.
=============== ====================================================
:returns: string
"""
out_location = os.path.dirname(location)
fc_name = os.path.basename(location)
df = geo._data
old_idx = df.index
df.reset_index(drop=True, inplace=True)
if geo.name is None:
raise ValueError("DataFrame must have geometry set.")
if validate and \
geo.validate(strict=True) == False:
raise ValueError(("Mixed geometry types detected, "
"cannot export to feature class."))
# sanitize
if sanitize_columns:
# logic
_sanitize_column_names(geo, inplace=True)
columns = df.columns.tolist()
for col in columns[:]:
if not isinstance(col, str):
df.rename(columns={col: str(col)}, inplace=True)
col = str(col)
if HASARCPY:
# 1. Create the Save Feature Class
#
columns = df.columns.tolist()
join_dummy = "AEIOUYAJC81Z"
columns.pop(columns.index(df.spatial.name))
dtypes = [(join_dummy, np.int64)]
if overwrite and arcpy.Exists(location):
arcpy.Delete_management(location)
elif overwrite == False and arcpy.Exists(location):
raise ValueError(('overwrite set to False, Cannot '
'overwrite the table. '))
notnull = geo._data[geo._name].notnull()
idx = geo._data[geo._name][notnull].first_valid_index()
sr = geo._data[geo._name][idx]['spatialReference']
gt = geo._data[geo._name][idx].geometry_type.upper()
null_geom = {
'point':
pd.io.json.dumps({
'x': None,
'y': None,
'spatialReference': sr
}),
'polyline':
pd.io.json.dumps({
'paths': [],
'spatialReference': sr
}),
'polygon':
pd.io.json.dumps({
'rings': [],
'spatialReference': sr
}),
'multipoint':
pd.io.json.dumps({
'points': [],
'spatialReference': sr
})
}
sr = geo._data[geo._name][idx].spatial_reference.as_arcpy
null_geom = null_geom[gt.lower()]
if has_m == True:
has_m = "ENABLED"
else:
has_m = None
if has_z == True:
has_z = "ENABLED"
else:
has_z = None
fc = arcpy.CreateFeatureclass_management(out_location,
spatial_reference=sr,
geometry_type=gt,
out_name=fc_name,
has_m=has_m,
has_z=has_z)[0]
# 2. Add the Fields and Data Types
oidfld = da.Describe(fc)['OIDFieldName']
for col in columns[:]:
if col.lower() in ['fid', 'oid', 'objectid']:
dtypes.append((col, np.int32))
elif df[col].dtype.name.startswith('datetime64[ns'):
dtypes.append((col, '<M8[us]'))
elif df[col].dtype.name == 'object':
try:
u = type(df[col][df[col].first_valid_index()])
except:
u = pd.unique(df[col].apply(type)).tolist()[0]
if issubclass(u, str):
mlen = df[col].str.len().max()
dtypes.append((col, '<U%s' % int(mlen)))
else:
try:
if df[col][idx] is None:
dtypes.append((col, '<U254'))
else:
dtypes.append((col, type(df[col][idx])))
except:
dtypes.append((col, '<U254'))
elif df[col].dtype.name == 'int64':
dtypes.append((col, np.int64))
elif df[col].dtype.name == 'bool':
dtypes.append((col, np.int32))
else:
dtypes.append((col, df[col].dtype.type))
array = np.array([], np.dtype(dtypes))
arcpy.da.ExtendTable(fc, oidfld, array, join_dummy, append_only=False)
# 3. Insert the Data
fields = arcpy.ListFields(fc)
icols = [fld.name for fld in fields \
if fld.type not in ['OID', 'Geometry'] and \
fld.name in df.columns] + ['SHAPE@JSON']
dfcols = [fld.name for fld in fields \
if fld.type not in ['OID', 'Geometry'] and\
fld.name in df.columns] + [df.spatial.name]
with da.InsertCursor(fc, icols) as irows:
dt_fld_idx = [irows.fields.index(col) for col in df.columns \
if df[col].dtype.name.startswith('datetime64[ns')]
def _insert_row(row):
row[-1] = pd.io.json.dumps(row[-1])
for idx in dt_fld_idx:
if isinstance(row[idx], type(pd.NaT)):
row[idx] = None
irows.insertRow(row)
q = df[geo._name].isna()
df.loc[q, 'SHAPE'] = null_geom # set null values to proper JSON
np.apply_along_axis(_insert_row, 1, df[dfcols].values)
df.loc[q, 'SHAPE'] = None # reset null values
df.set_index(old_idx)
return fc
elif HASPYSHP:
if fc_name.endswith('.shp') == False:
fc_name = "%s.shp" % fc_name
if SHPVERSION < [2]:
res = _pyshp_to_shapefile(df=df,
out_path=out_location,
out_name=fc_name)
df.set_index(old_idx)
return res
else:
res = _pyshp2(df=df, out_path=out_location, out_name=fc_name)
df.set_index(old_idx)
return res
elif HASARCPY == False and HASPYSHP == False:
raise Exception(
("Cannot Export the data without ArcPy or PyShp modules."
" Please install them and try again."))
else:
df.set_index(old_idx)
return None
def to_table(geo, location, overwrite=True):
"""
Exports a geo enabled dataframe to a table.
=========================== ====================================================================
**Argument** **Description**
--------------------------- --------------------------------------------------------------------
location Required string. The output of the table.
--------------------------- --------------------------------------------------------------------
overwrite Optional Boolean. If True and if the table exists, it will be
deleted and overwritten. This is default. If False, the table and
the table exists, and exception will be raised.
=========================== ====================================================================
:returns: String
"""
out_location = os.path.dirname(location)
fc_name = os.path.basename(location)
df = geo._data
if location.lower().find('.csv') > -1:
geo._df.to_csv(location)
return location
elif HASARCPY:
columns = df.columns.tolist()
join_dummy = "AEIOUYAJC81Z"
try:
columns.pop(columns.index(df.spatial.name))
except:
pass
dtypes = [(join_dummy, np.int64)]
if overwrite and arcpy.Exists(location):
arcpy.Delete_management(location)
elif overwrite == False and arcpy.Exists(location):
raise ValueError(('overwrite set to False, Cannot '
'overwrite the table. '))
fc = arcpy.CreateTable_management(out_path=out_location,
out_name=fc_name)[0]
# 2. Add the Fields and Data Types
#
oidfld = da.Describe(fc)['OIDFieldName']
for col in columns[:]:
if col.lower() in ['fid', 'oid', 'objectid']:
dtypes.append((col, np.int32))
elif df[col].dtype.name == 'datetime64[ns]':
dtypes.append((col, '<M8[us]'))
elif df[col].dtype.name == 'object':
try:
u = type(df[col][df[col].first_valid_index()])
except:
u = pd.unique(df[col].apply(type)).tolist()[0]
if issubclass(u, str):
mlen = df[col].str.len().max()
dtypes.append((col, '<U%s' % int(mlen)))
else:
try:
dtypes.append((col, type(df[col][0])))
except:
dtypes.append((col, '<U254'))
elif df[col].dtype.name == 'int64':
dtypes.append((col, np.int64))
elif df[col].dtype.name == 'bool':
dtypes.append((col, np.int32))
else:
dtypes.append((col, df[col].dtype.type))
array = np.array([], np.dtype(dtypes))
arcpy.da.ExtendTable(fc, oidfld, array, join_dummy, append_only=False)
# 3. Insert the Data
#
fields = arcpy.ListFields(fc)
icols = [fld.name for fld in fields \
if fld.type not in ['OID', 'Geometry'] and \
fld.name in df.columns]
dfcols = [fld.name for fld in fields \
if fld.type not in ['OID', 'Geometry'] and\
fld.name in df.columns]
with da.InsertCursor(fc, icols) as irows:
for idx, row in df[dfcols].iterrows():
try:
irows.insertRow(row.tolist())
except:
print("row %s could not be inserted." % idx)
return fc
return