def __init__(self, filename, year=None, encoding='GBK'):
self.cgss_db = CgssDatabase()
self.stata_object = Stata(filename, encoding=encoding)
self.stata_file = self.stata_object.stata_file
self.year = int(year)
# 变量-标签字典
self.variables_labels_dict = dict(
zip(self.stata_object.variables, self.stata_object.variables_labels))
# 变量-值标签字典
self.variables_values_labels_dict = dict(
zip(self.stata_object.variables, self.stata_object.variables_values_labels))
# 值标签字典
self.values_labels_dict = self.values_labels
class CgssStataSheet:
"""类CgssStataSheet用来读写cgss.dta（stata格式）的数据文件
:param str filename: 想要读写的文件名
:param int year: 年份
:param str encoding: 编码方式，默认是'GBK'
:return: 无返回值
:var list variables: 变量列表
:var list variables_labels: 变量标签列表
:var list variables_values_labels: 变量的值标签
:var list values_labels: 值标签的内容
:var dict variables_labels_dict: 变量-标签字典
:var dict variables_values_labels_dict: 变量-值标签字典
:var dict values_labels_dict: 值-标签字典
"""
def __init__(self, filename, year=None, encoding='GBK'):
self.cgss_db = CgssDatabase()
self.stata_object = Stata(filename, encoding=encoding)
self.stata_file = self.stata_object.stata_file
self.year = int(year)
# 变量-标签字典
self.variables_labels_dict = dict(
zip(self.stata_object.variables, self.stata_object.variables_labels))
# 变量-值标签字典
self.variables_values_labels_dict = dict(
zip(self.stata_object.variables, self.stata_object.variables_values_labels))
# 值标签字典
self.values_labels_dict = self.values_labels
def insert(self):
"""insert方法用来插入数据到Mongodb数据库集合
"""
# 读入stata文件数据
data = self.stata_object.read()
# 数据行数
data_rows = data.shape[0]
print(data_rows)
# 读取每一行数据（产生一条记录）
for i in range(data_rows):
# 每一条记录用字典记录，存储在record变量中
record = dict({'year': self.year})
# 读取一条记录
row_data = data.iloc[i]
j = 0
# 包装记录，即记录变量数据在字典record中
# ind是行数据的索引（即变量）
for ind in row_data.index:
j = j + 1
# 初始化值标签
value_label = None
# 变量-值标签变量
variable_value_labels = self.variables_values_labels_dict[ind]
# 值标签字典
value_labels = self.values_labels_dict.get(
variable_value_labels, None)
# 变量的值
value = row_data[ind]
# 如果变量的值是数值型
if isinstance(value, (np.float32, np.float64, np.integer)):
if np.isnan(value):
value = None
else:
if value_labels is None:
value = None
else:
if np.equal(value, np.int(value)):
value = int(value)
value_label = value_labels.get(
np.int(value), None)
else:
value = float(value)
elif isinstance(value, str):
if re.match('^\s*$', value) is not None:
value = None
elif isinstance(value, (pd.tslib.Timestamp, datetime)):
if isinstance(value, pd.tslib.NaTType):
value = None
else:
value = re.split(' ', str(value))[0]
else:
print(ind)
print(value, type(value))
raise TypeError
record[ind] = {'label': self.variables_labels_dict[ind], 'serial_number': j,
'value': {
'value': value,
'label': value_label
}}
print(i)
print('record', record)
self.cgss_db.collection.insert(record)
@property
def values_labels(self):
labels = self.stata_object.values_labels
new_labels_dict = dict()
for var in labels:
var_label = list(labels[var].items())
label_tuple = [(np.uint(item[0]).astype(int), item[1])
for item in var_label]
new_labels_dict[var] = dict(label_tuple)
return new_labels_dict
