def gen_data_simple_schema(self, data_path, partition_date, num_rows,
file_format):
"""
Input
- data_path: path where the partition will be created (string)
- partition_date: partition date to be created (date)
- num_rows: number of rows to be generated (integer)
- file_format: format of file to be generated (parquet or avro)
This function creates a data sample with a simple schema
"""
person = Person('en')
# Create a simple schema
schema_df = StructType([
StructField('identifier', StringType(), True),
StructField('first_name', StringType(), True),
StructField('last_name', StringType(), True),
StructField('occupation', StringType(), True),
StructField('age', IntegerType(), True),
StructField('date', DateType(), True)
])
# generate data
for _ in range(num_rows):
df_temp = self.spark.createDataFrame([[
person.identifier(),
person.first_name(),
person.last_name(),
person.occupation(),
person.age(), partition_date
]], schema_df)
try:
df = df.union(df_temp)
except:
df = df_temp
df.coalesce(1).write.partitionBy('date').mode('overwrite').format(
file_format).save(data_path)
print('Partition created: {data_path}/date={date}'.format(
data_path=data_path, date=partition_date))
print('# Rows:', df.count())
print('Schema:')
df.printSchema()
print('\n')
return
def getting_started_example():
generic = Generic()
#generic = Generic(locales.EN)
print('Month =', generic.datetime.month())
print('Datetime =',
generic.datetime.datetime(start=1900, end=2035,
timezone=None)) # Type: datetime.datetime.
print('IMEI =', generic.code.imei())
print('Fruit =', generic.food.fruit())
print('RNA =', generic.science.rna_sequence())
print('Word =', generic.text.word())
with generic.text.override_locale(locales.FR):
print('Word =', generic.text.word())
print('Word =', generic.text.word())
generic = Generic('en')
generic.add_provider(USASpecProvider)
print('SSN =', generic.usa_provider.ssn())
#print('CPF =', generic.usa_provider.cpf()) # AttributeError: 'USASpecProvider' object has no attribute 'cpf'.
generic = Generic('pt-br')
#generic = Generic(locales.PT_BR)
generic.add_provider(BrazilSpecProvider)
#print('SSN =', generic.brazil_provider.ssn()) # AttributeError: 'BrazilSpecProvider' object has no attribute 'ssn'.
print('CPF =', generic.brazil_provider.cpf())
#--------------------
numbers = Numbers()
print('Numbers =', numbers.between()) # Type: int.
print('Numbers =', numbers.between(10, 10000000000000000)) # Type: int.
#--------------------
person = Person(locales.KO)
print('Full name =', person.full_name(gender=Gender.FEMALE))
print('Full name =', person.full_name(gender=Gender.MALE, reverse=True))
with person.override_locale(locales.RU):
print('Full name =', person.full_name())
print('Full name =', person.full_name())
print('Telephone =', person.telephone())
print('Telephone =', person.telephone(mask='(###)-###-####'))
print('Identifier =', person.identifier())
print('Identifier =', person.identifier(mask='######-#######'))
#--------------------
de = Address('de')
ru = Address('ru')
print('Region =', de.region())
print('Federal subject =', ru.federal_subject())
print('Address =', de.address())
print('Address =', ru.address())
ko = Address('ko')
print('Address =', ko.province(), ko.city(), ko.address())
print('Zip code =', ko.zip_code())
#--------------------
business = Business('ko')
#print('Price =', business.price(minimum=1.0, maximum=1000000000.0)) # Type: str.
#print('Price =', business.price(minimum=1.0, maximum=1000000000.0)[:-2]) # Type: str.
print('Price =',
business.price(minimum=1.0, maximum=1000000000.0)[:-5]) # Type: str.
#--------------------
payment = Payment()
print('Credit card =',
payment.credit_card_number(card_type=None)) # Type: str.
def person(
cls,
*,
locale=Locales.EN,
qualification=None,
age=None,
blood_type=None,
email=None,
first_name=None,
last_name=None,
gender=None,
height=None,
id=None,
language=None,
nationality=None,
occupation=None,
phone=None,
title=None,
university=None,
weight=None,
work_experience=None,
):
'''
Create an Person Data Entity object.
All individual fields are automatically randomly generated based on locale. If provided, the corresponding values are overriden.
Note:
All individual fields are randomly generated. Don't expect correct correlation e.g. correct postal code for the generated city.
Keyword Arguments:
locale: Approprite Random.locale.<local_name> object. Default is Random.locale.EN
qualification: Educational Qualification
age: Age
blood_type: Blood type
email: Email address
first_name: First name
last_name: Last name
gender: Gender
height: Height
id: Identifier
language: Language
nationality: Nationality
occupation: Occupation
phone: Phone number
title: Title
university: University
weight: Weight
work_experience: Work Experience
'''
person = Person(locale=locale)
from arjuna.engine.data.entity.person import Person as ArjPerson
first_name = first_name is not None and first_name or person.first_name(
)
last_name = last_name is not None and last_name or person.last_name()
return ArjPerson(
qualification=qualification is not None and qualification
or person.academic_degree(),
age=age is not None and age or person.age(),
blood_type=blood_type is not None and blood_type
or person.blood_type(),
email=email is not None and email or person.email(),
first_name=first_name,
last_name=last_name,
name=first_name + " " + last_name,
gender=gender is not None and gender or person.gender(),
height=height is not None and height or person.height(),
id=id is not None and id or person.identifier(),
language=language is not None and language or person.language(),
nationality=nationality is not None and nationality
or person.nationality(),
occupation=occupation is not None and occupation
or person.occupation(),
phone=phone is not None and phone or person.telephone(),
title=title is not None and title or person.title(),
university=university is not None and university
or person.university(),
weight=weight is not None and weight or person.weight(),
work_experience=work_experience is not None and work_experience
or person.work_experience(),
)
def gen_data_remove_column(self, data_path, partition_date, num_rows,
file_format):
"""
Input
- data_path: path where the partition will be created (string)
- partition_date: partition date to be created (date)
- num_rows: number of rows to be generated (integer)
- file_format: format of file to be generated (parquet or avro)
This function creates a data sample removing some columns
"""
person = Person('en')
address = Address('en')
schema_street = StructType([
StructField('street_name', StringType(), True)
# StructField('lat', FloatType(), True), #column removed
# StructField('long', FloatType(), True) #column removed
])
schema_address_details = StructType([
StructField('street', schema_street, True),
StructField('number', IntegerType(), True)
])
schema_address = StructType([
StructField('address_details', schema_address_details, True),
StructField('city', StringType(), True),
StructField('country', StringType(), True),
# StructField('country_code', StringType(), True), #column removed
StructField('state', StringType(), True),
StructField('postal_code', IntegerType(), True)
])
schema_df = StructType([
StructField('identifier', StringType(), True),
StructField('first_name', StringType(), True),
StructField('last_name', StringType(), True),
StructField('occupation', StringType(), True),
StructField('age', IntegerType(), True),
StructField('address', schema_address, True),
# StructField('title_name', StringType(), True), #column removed
StructField('date', DateType(), True)
])
for _ in range(num_rows):
df_temp = self.spark.createDataFrame(
[[
person.identifier(),
person.first_name(),
person.last_name(),
person.occupation(),
person.age(),
[
[
[
address.street_name()
#float(address.latitude()),
#float(address.longitude())
],
int(address.street_number())
],
address.city(),
address.country(),
#address.country_code(),
address.state(),
int(address.postal_code())
],
#person.title(),
partition_date
]],
schema_df)
try:
df = df.union(df_temp)
except:
df = df_temp
df.coalesce(1).write.partitionBy('date').mode('overwrite').format(
file_format).save(data_path)
print('Partition created: {data_path}/date={date}'.format(
data_path=data_path, date=partition_date))
print('# Rows:', df.count())
print('Schema:')
df.printSchema()
print('\n')
return
