def _build_normalization_factor(self,
select_from,
select_from_as="_visit_count"):
# compute normalization factor
normalization_factor = SelectQuery()
normalization_factor.select_from(select_from, alias=select_from_as)
self._select_columns_list(
normalization_factor,
column_names=self.similarity_computation_columns +
self.filtering_columns,
table_name=select_from_as,
)
rating_column = (Column(self.dku_config.ratings_column_name).minus(
Column(self.RATING_AVERAGE)) if self.use_explicit else Constant(1))
normalization_factor.select(
self._get_normalization_factor_formula(Column(self.based_column),
rating_column),
alias=self.NORMALIZATION_FACTOR_AS,
)
self.similarity_computation_columns += [self.NORMALIZATION_FACTOR_AS]
# keep only items and users with enough visits
normalization_factor.where(
Column(self.NB_VISIT_USER_AS, table_name=select_from_as).ge(
Constant(self.dku_config.user_visit_threshold)))
normalization_factor.where(
Column(self.NB_VISIT_ITEM_AS, table_name=select_from_as).ge(
Constant(self.dku_config.item_visit_threshold)))
return normalization_factor
def _build_cf_scores_without_null(self, select_from, select_from_as="_all_cf_scores_to_filter"):
null_scores_filtered = SelectQuery()
null_scores_filtered.select_from(select_from, alias=select_from_as)
columns_to_select = self.sample_keys + self.dku_config.score_column_names
self._select_columns_list(select_query=null_scores_filtered, column_names=columns_to_select)
self._or_condition_columns_list(
null_scores_filtered, self.dku_config.score_column_names, lambda x: x.is_not_null()
)
return null_scores_filtered
def _build_all_cf_scores(
self,
select_from,
samples_for_training,
samples_for_scores=None,
select_from_as="_all_cf_scores",
samples_for_training_as="_samples_for_training",
samples_for_scores_as="_samples_for_scores",
):
all_cf_scores = SelectQuery()
all_cf_scores.select_from(select_from, alias=select_from_as)
self._left_join_samples(
all_cf_scores, select_from_as, samples_for_training, samples_for_training_as, self.sample_keys
)
all_cf_scores.select(Column(self.IS_TRAINING_SAMPLE, table_name=samples_for_training_as))
if samples_for_scores:
self._left_join_samples(
all_cf_scores, select_from_as, samples_for_scores, samples_for_scores_as, self.sample_keys
)
all_cf_scores.select(Column(self.IS_SCORE_SAMPLE, table_name=samples_for_scores_as))
columns_to_select = self.sample_keys + self.dku_config.score_column_names
self._select_columns_list(all_cf_scores, columns_to_select, table_name=select_from_as)
return all_cf_scores
def _build_samples_with_all_infos(inner_select_from, join_with, inner_select_from_as=ALL_INFOS_TABLE_NAME):
samples_with_all_infos = SelectQuery()
samples_with_all_infos.select_from(inner_select_from, alias=inner_select_from_as)
columns_to_select = self.sample_keys + self.dku_config.score_column_names + [constants.TARGET_COLUMN_NAME]
self._select_columns_list(samples_with_all_infos, columns_to_select, table_name=inner_select_from_as)
row_number_expression = (
Expression()
.rowNumber()
.over(
Window(
partition_by=[
Column(self.dku_config.users_column_name, table_name=inner_select_from_as),
Column(constants.TARGET_COLUMN_NAME, table_name=inner_select_from_as),
],
order_by=[Column(constants.TARGET_COLUMN_NAME, table_name=inner_select_from_as)],
order_types=["DESC"],
mode=None,
)
)
)
samples_with_all_infos.select(row_number_expression, alias=self.ROW_NUMBER_AS)
samples_with_all_infos.select(Column(NB_POSITIVE_PER_USER, table_name=ONLY_POSITIVE_TABLE_NAME))
self._left_join_samples(
left_select_query=samples_with_all_infos,
left_table_name=inner_select_from_as,
right_select_query=join_with,
right_table_name=ONLY_POSITIVE_TABLE_NAME,
keys=[self.dku_config.users_column_name],
)
return samples_with_all_infos
def make_most_recent_timestamp_query(timestamp_column, dataset):
is_hdfs = 'hiveTableName' in dataset.get_config().get('params').keys()
query = SelectQuery()
query.select(Column(timestamp_column).max(), alias='most_recent_timestamp')
if is_hdfs:
query.select_from('_'.join(dataset.name.split('.')))
else:
query.select_from(dataset)
dialect_handler = dialectHandler(dataset)
return dialect_handler.convertToSQL(
query) #toSQL(query, dialect='Hive')#dataset=dataset)
def make_max_time_interval_query(timestamp_column, resolved_ref_date, dataset):
is_hdfs = 'hiveTableName' in dataset.get_config().get('params').keys()
max_time_interval = Constant(resolved_ref_date).minus(
Column(timestamp_column)).extract(TimeUnit.DAY).max()
query = SelectQuery()
query.select(max_time_interval, alias="max_time_interval") #TODO naming
if is_hdfs:
query.select_from('_'.join(dataset.name.split('.')))
else:
query.select_from(dataset)
dialect_handler = dialectHandler(dataset)
return dialect_handler.convertToSQL(query) #toSQL(query, dataset=dataset)
def _build_all_cf_scores_with_target(self, select_from, select_from_as="_all_cf_scores_with_target"):
all_cf_scores_with_target = SelectQuery()
all_cf_scores_with_target.select_from(select_from, alias=select_from_as)
columns_to_select = self.sample_keys + self.dku_config.score_column_names
self._select_columns_list(select_query=all_cf_scores_with_target, column_names=columns_to_select)
all_cf_scores_with_target.select(
Column(self.IS_TRAINING_SAMPLE).coalesce(0).cast("int"), alias=constants.TARGET_COLUMN_NAME
)
if self.has_historical_data:
all_cf_scores_with_target.select(
Column(self.IS_SCORE_SAMPLE).coalesce(0).cast("int"), alias=self.SCORE_SAMPLE
)
return all_cf_scores_with_target
def _encoding_feature_name(self, dataset):
is_hdfs = 'hiveTableName' in dataset.get_config().get('params').keys()
dataset_schema = dataset.read_schema()
col_list = []
new_col_alias = []
for col_index, col_info in enumerate(dataset_schema):
col_name = col_info.get('name')
col_list.append(Column(col_name))
col_name_mapping = self.params.feature_name_mapping.get(col_name)
if col_name_mapping:
col_alias = '{}_{}'.format(col_name_mapping, col_index)
else:
col_alias = col_name
new_col_alias.append(col_alias)
query_to_rename = SelectQuery().select(col_list, new_col_alias)
if is_hdfs:
query_to_rename.select_from('_'.join(dataset.name.split('.')))
else:
query_to_rename.select_from(dataset)
dialect_handler = dialectHandler(dataset)
dialect_handler.get_executor().exec_recipe_fragment(
dataset, query=dialect_handler.convertToSQL(
query_to_rename)) #toSQL(query_to_rename, dataset=dataset))
def _rename_table(self, to_rename, renaming_mapping):
renamed_table = SelectQuery()
renamed_table.select_from(to_rename, alias="_renamed")
for input_column, renamed_column in renaming_mapping.items():
renamed_table.select(Column(input_column, table_name="_renamed"),
alias=renamed_column)
return renamed_table
def _cast_table(self, to_cast, cast_mapping, alias):
cast_table = SelectQuery()
cast_table.select_from(to_cast, alias=alias)
for input_column, target_type in cast_mapping.items():
cast_table.select(Column(input_column,
table_name=alias).cast(target_type),
alias=input_column)
return cast_table
def _build_remove_historical_samples(self, select_from, select_from_as="_remove_negative_samples_seen"):
historical_negative_samples_removed = SelectQuery()
historical_negative_samples_removed.select_from(select_from, alias=select_from_as)
columns_to_select = self.sample_keys + [constants.TARGET_COLUMN_NAME] + self.dku_config.score_column_names
self._select_columns_list(select_query=historical_negative_samples_removed, column_names=columns_to_select)
unseen_samples_condition = (
Column(constants.TARGET_COLUMN_NAME).eq(Constant(1)).or_(Column(self.SCORE_SAMPLE).eq(Constant(0)))
)
historical_negative_samples_removed.where(unseen_samples_condition)
return historical_negative_samples_removed
def _build_filtered_samples(inner_select_from, inner_select_from_as):
ratio = float(self.dku_config.negative_samples_percentage / 100.0)
filtered_samples = SelectQuery()
filtered_samples.select_from(inner_select_from, inner_select_from_as)
columns_to_select = self.sample_keys + self.dku_config.score_column_names + [constants.TARGET_COLUMN_NAME]
self._select_columns_list(filtered_samples, columns_to_select)
nb_negative_threshold_expr = (
Column(NB_POSITIVE_PER_USER, table_name=select_from_as)
.times(Constant(ratio))
.div(Constant(1).minus(Constant(ratio)))
.ceil()
)
filtered_samples.where(
Column(constants.TARGET_COLUMN_NAME, table_name=select_from_as)
.eq(Constant(1))
.or_(Column(self.ROW_NUMBER_AS, table_name=select_from_as).le(nb_negative_threshold_expr))
)
return filtered_samples
def _build_unordered_similarity(
self,
select_from,
left_select_from_as="_left_ordered_similarity",
right_select_from_as="_right_ordered_similarity",
with_clause_as="_with_clause_ordered_similarity",
):
"""Retrieve both pairs (when col_1 < col_2 and col_1 > col_2) from the ordered similarity table"""
similarity = SelectQuery()
if self.supports_with_clause:
similarity.with_cte(select_from, alias=with_clause_as)
select_from = with_clause_as
similarity.select_from(select_from, alias=left_select_from_as)
join_condition = Constant(1).eq_null_unsafe(Constant(0))
similarity.join(select_from,
JoinTypes.FULL,
join_condition,
alias=right_select_from_as)
similarity.select(
Column(f"{self.based_column}_1",
table_name=left_select_from_as).coalesce(
Column(f"{self.based_column}_2",
table_name=right_select_from_as)),
alias=f"{self.based_column}_1",
)
similarity.select(
Column(f"{self.based_column}_2",
table_name=left_select_from_as).coalesce(
Column(f"{self.based_column}_1",
table_name=right_select_from_as)),
alias=f"{self.based_column}_2",
)
similarity.select(
Column("similarity", table_name=left_select_from_as).coalesce(
Column("similarity", table_name=right_select_from_as)),
alias=constants.SIMILARITY_COLUMN_NAME,
)
return similarity
def make_full_transform_query(aggregation_queries,
dataset,
aggregation_params,
transform_params,
encoding_feature=False):
is_hdfs = 'hiveTableName' in dataset.get_config().get('params').keys()
inner = SelectQuery()
if is_hdfs:
inner.select_from('_'.join(dataset.name.split('.')))
else:
inner.select_from(dataset)
if aggregation_params.is_rolling_window():
inner.select(Column('*'))
else:
inner.distinct() #TODO why?! -> avoid dupplicate
for key in aggregation_params.get_effective_keys():
inner.select(Column(key))
prefilter = _make_prefilter(aggregation_params, transform_params)
inner.where(prefilter)
outer = SelectQuery()
outer.select_from(inner, alias='inner')
if aggregation_params.is_rolling_window():
outer.select(Column('*', 'inner'))
else:
for col in aggregation_params.get_effective_keys(): #+ feature_names:
outer.select(Column(col, 'inner'))
reverse_mapping_dict = {}
for idx, agg_query in enumerate(aggregation_queries):
agg_query.alias(
agg_query.get_alias()
or 'cte_' + str(idx)) #TODO remove, make sure they have ids
outer.with_cte(agg_query)
join_cond = Expression()
for key in aggregation_params.get_effective_keys():
join_cond = join_cond.and_(
Column(key, 'inner').eq_null_unsafe(
Column(key, agg_query.get_alias())))
outer.join(agg_query.get_alias(), JoinTypes.LEFT, join_cond)
for idx2, col in enumerate(agg_query.get_columns_alias()):
if encoding_feature:
if aggregation_params.feature_name_mapping.get(col):
new_alias = '{}_{}_{}'.format(
aggregation_params.feature_name_mapping.get(col), idx,
idx2)
outer.select(Column(col, agg_query.get_alias()), new_alias)
reverse_mapping_dict[new_alias] = col
else:
outer.select(Column(col, agg_query.get_alias()))
return dialectHandler(dataset).convertToSQL(outer), reverse_mapping_dict
def make_distinct_values_query(column, dataset):
is_hdfs = 'hiveTableName' in dataset.get_config().get('params').keys()
column = Column(column)
layer_1 = SelectQuery()
layer_1.select(column)
layer_1.select(Column('*').count(), 'count')
if is_hdfs:
layer_1.select_from('_'.join(dataset.name.split('.')))
else:
layer_1.select_from(dataset)
layer_1.group_by(column)
count = Column('count')
layer_2 = SelectQuery()
layer_2.select(column)
layer_2.select(
count.div(count.sum().over(Window())).times(100), 'distribution')
layer_2.select_from(layer_1, alias='layer_1')
dialect_handler = dialectHandler(dataset)
return dialect_handler.convertToSQL(
layer_2) #toSQL(query, dialect='Hive')#dataset=dataset)
def _build_samples_for_training(self, select_from, select_from_as="_samples_for_training"):
samples_for_training = SelectQuery()
samples_for_training.select_from(select_from, alias=select_from_as)
self._select_columns_list(samples_for_training, self.sample_keys, table_name=select_from_as)
samples_for_training.select(Constant(1), alias=self.IS_TRAINING_SAMPLE)
return samples_for_training
def _build_samples_for_scoring(self, select_from, select_from_as="_samples_for_scores"):
samples_for_scores = SelectQuery()
samples_for_scores.select_from(select_from, alias=select_from_as)
self._select_columns_list(samples_for_scores, self.sample_keys, table_name=select_from_as)
samples_for_scores.select(Constant(1), alias=self.IS_SCORE_SAMPLE)
return samples_for_scores
# It does not replace proper documentation, tutos and samples but for now it will do the trick
# -------------------------------------------------------------------------------- NOTEBOOK-CELL: CODE
import dataiku.sql as sql
import json
from dataiku.sql import JoinTypes, Expression, Column, Constant, InlineSQL, SelectQuery, Table, Dialects, toSQL
# -------------------------------------------------------------------------------- NOTEBOOK-CELL: CODE
# Expression: 1 * 2 * 3
eb = Constant(1).times(Constant(2), Constant(3))
toSQL(eb, Dialects.MYSQL)
# -------------------------------------------------------------------------------- NOTEBOOK-CELL: CODE
# Simple select
sb = SelectQuery()
sb.select_from(Table('myTable'))
toSQL(sb, Dialects.MYSQL)
# -------------------------------------------------------------------------------- NOTEBOOK-CELL: CODE
# join in fluent style
sb = SelectQuery() \
.select(Constant('Hello')) \
.select_from('t1') \
.join('t2', JoinTypes.INNER, Column('c', 't1').eq(Column('c', 't2')))
toSQL(sb, Dialects.MYSQL)
# -------------------------------------------------------------------------------- NOTEBOOK-CELL: CODE
# Group by and limit
def _build_samples_with_only_positives(inner_select_from, inner_select_from_as=ONLY_POSITIVE_TABLE_NAME):
samples_with_only_positives = SelectQuery()
samples_with_only_positives.select_from(inner_select_from, inner_select_from_as)
samples_with_only_positives.select(Column(self.dku_config.users_column_name))
samples_with_only_positives.select(Column("*").count(), alias=NB_POSITIVE_PER_USER)
samples_with_only_positives.where(Column(constants.TARGET_COLUMN_NAME).eq(Constant(1)))
samples_with_only_positives.group_by(Column(self.dku_config.users_column_name))
return samples_with_only_positives
