def valid_redemption(self, grouping_set, column_set, measure):
df_correct = self.base_redemption_df.filter(
col('contact_stage_code') == 'EXP'
).select(
self.config_dict['identity_type_code'],
'offer_code'
).dropDuplicates()
df = self.base_redemption_df.filter(
col('contact_stage_code') == 'RDM'
).join(
df_correct,
[self.config_dict['identity_type_code'], 'offer_code'],
'inner'
)
df_final = utils.count(
self.sqlContext,
df,
grouping_set,
column_set,
measure,
self.config_dict['identity_type_code']
)
return df_final
def digital_trigered(self, grouping_set, column_set, measure):
df = self.base_redemption_df.filter(
(lower(col('channel_code')).like('%digital%'))
& (col('contact_stage_code') == 'ACT')
)
df_final = utils.count(
self.sqlContext,
df,
grouping_set,
column_set,
measure,
'offer_code'
)
return df_final
def number_of_redemptions(self, grouping_set, column_set, measure):
if not self.df_redemptions.head(1):
df_offer = utils.count(
self.sqlContext,
self.df_redeem,
grouping_set,
column_set,
measure,
'offer_code'
)
self.df_redemptions = df_offer
else:
df_offer = self.df_redemptions
return df_offer
def mis_redemptions(self, grouping_set, column_set, measure):
df_subtract = self.df_redeem.select(
'prod_code', 'offer_code').dropDuplicates().subtract(
self.df_dict('table3').select('prod_code',
'offer_code').dropDuplicates())
df_subtract_prsn = df_subtract.join(self.df_redeem,
['prod_code', 'offer_code'])
# since f.count() don't consider null customers
df_subtract_prsn = df_subtract_prsn.fillna('null', ['prsn_code'])
df_final = utils.count(self.sqlContext, df_subtract_prsn, grouping_set,
column_set, measure,
self.config_dict['identity_type_code'])
return df_final
def redemption_rate(self, grouping_set, column_set, measure):
redemeed = self.base_redemption_df.filter(
col('contact_stage_code') == 'RDM'
)
# print 'redemeed'
# redemeed.cache()
# print redemeed.count()
mailed = self.base_redemption_df.filter(
col('contact_stage_code') == 'EXP'
)
# print 'mailed'
# mailed.cache()
# print mailed.count()
redem_coupon = utils.count(
self.sqlContext,
redemeed,
grouping_set,
column_set,
'redem_coupon',
'offer_code'
)
# print 'redem_coupon'
# redem_coupon.cache()
# print redem_coupon.count()
mailed_coupon = utils.count(
self.sqlContext,
mailed,
grouping_set,
column_set,
'mailed_coupon',
'offer_code'
)
# print 'mailed_coupon'
# mailed_coupon.cache()
# print mailed_coupon.count()
group_set = column_set + ['grouping_level']
df_redem_mailed = redem_coupon.join(mailed_coupon, group_set)
# print 'df_redem_mailed'
# df_redem_mailed.cache()
# print df_redem_mailed.count()
df_final = df_redem_mailed.withColumn(
measure,
df_redem_mailed.redem_coupon/df_redem_mailed.mailed_coupon
)
# print 'df_final'
# df_final.cache()
# print df_final.count()
df_redemption_rate = df_final.drop('redem_coupon')
df_redemption_rate = df_redemption_rate.drop('mailed_coupon')
df_redemption_rate = df_redemption_rate.withColumn(
measure,
df_redemption_rate[measure].cast(StringType())
)
return df_redemption_rate
def coupons_allocated(self, grouping_set, column_set, measure):
df = self.base_allocation_df.filter(col('contact_stage_code') == 'DLV')
coupon_allocated_df = utils.count(self.sqlContext, df, grouping_set,
column_set, measure, 'offer_code')
return coupon_allocated_df
def buy_in_category_but_not_product(self, grouping_set, column_set, measure):
exp_df = self.detail_offer_prod
# .withColumnRenamed('prod_code', 'featured_prod_code')
# print 'exp_df'
# exp_df.cache()
# print exp_df.count()
# exp_df.show()
prod_dim_df = self.df_dict('prod_dim')
# prod_dim_df.cache()
# prod_dim_df.show()
exp_cat_df = exp_df.join(
prod_dim_df,
'prod_code'
'left_outer'
)
exp_cat_df = exp_cat_df.withColumnRenamed('prod_code', 'featured_prod_code')
# print 'exp_cat_df'
# exp_cat_df.cache()
# print exp_cat_df.count()
# exp_cat_df.show()
pre_prod_df = self.df_dict('pre_period').join(
self.df_dict('prod_dim'),
'prod_code',
'left_outer'
)
# print 'pre_prod_df'
# pre_prod_df.cache()
# print pre_prod_df.count()
# pre_prod_df.show()
post_prod_df = self.df_dict('post_period').join(
self.df_dict('prod_dim'),
'prod_code',
'left_outer'
)
# print 'post_prod_df'
# post_prod_df.cache()
# print post_prod_df.count()
# post_prod_df.show()
pre_post_df = pre_prod_df.select(
self.config_dict['identity_type_code'],
'prod_hier_l20_code'
).intersect(
post_prod_df.select(
self.config_dict['identity_type_code'],
'prod_hier_l20_code'
)
)
# print 'pre_post_df'
# pre_post_df.cache()
# print pre_post_df.count()
# pre_post_df.show()
pre_post_exp_cat_df = exp_cat_df.join(
pre_post_df,
[self.config_dict['identity_type_code'], 'prod_hier_l20_code']
)
# print 'pre_post_exp_cat_df'
# pre_post_exp_cat_df.cache()
# print pre_post_exp_cat_df.count()
# pre_post_exp_cat_df.show()
df = pre_post_exp_cat_df.filter(
pre_post_exp_cat_df.featured_prod_code != pre_post_exp_cat_df.prod_code
)
# print 'df'
# df.cache()
# print df.count()
# df.show()
# .drop('channel_code')
buy_in_category_but_not_product_df = utils.count(
self.sqlContext,
df,
grouping_set,
column_set,
measure,
self.config_dict['identity_type_code']
)
return buy_in_category_but_not_product_df