def up(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name)
table = migration.client.get_table(dataset.table(table_name))
orig_schema = table.schema
new_schema = orig_schema.copy()
if table.schema[-1].name == 'client_paid_rank_level':
logging.warning(
f'rank levels already added to {table_name} in {dataset_name} dataset!'
)
return dataset
new_schema.extend([
bigquery.SchemaField('lifetime_rank_level', 'INTEGER'),
bigquery.SchemaField('paid_rank_level', 'INTEGER'),
bigquery.SchemaField('client_lifetime_rank_level', 'INTEGER'),
bigquery.SchemaField('client_paid_rank_level', 'INTEGER')
])
table.schema = new_schema
migration.client.update_table(table, ['schema'])
return dataset
def up(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name)
table = migration.client.get_table(dataset.table(table_name))
if table.schema[1].mode == 'NULLABLE':
logging.warning(
'leo_eid is already set to NULLABLE. Cannot do it again')
return dataset
new_schema = []
for f in table.schema:
if f.name == 'leo_eid':
new_schema.append(
bigquery.SchemaField('leo_eid', 'STRING', mode='NULLABLE'))
else:
new_schema.append(
bigquery.SchemaField(f.name, f.field_type, mode=f.mode))
table.schema = new_schema
migration.client.update_table(table, ['schema'])
return dataset
def up(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name)
table = migration.client.get_table(dataset.table(table_name))
new_schema = []
for f in table.schema:
if f.name == 'client_paid_rank' or f.name == 'paid_rank':
if f.mode == 'NULLABLE':
logging.warning(
'paid_rank fields are already set to NULLABLE. Cannot do it again'
)
return dataset
new_schema.append(
bigquery.SchemaField(f.name, f.field_type, 'NULLABLE'))
else:
new_schema.append(
bigquery.SchemaField(f.name, f.field_type, mode=f.mode))
table.schema = new_schema
migration.client.update_table(table, ['schema'])
return dataset
def up(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(
dataset_name
) # use me if you are NOT creating a new dataset. -- ndg 2/5/20
clusters = migration.default_clustering_fields
del clusters[0]
clusters.insert(0, 'client_wrench_id:STRING')
schema = [
bigquery.SchemaField('entity_id', 'STRING', mode='REQUIRED'),
bigquery.SchemaField("tree_user_id", "INTEGER", mode="REQUIRED"),
bigquery.SchemaField("prediction", "STRING", mode="REQUIRED"),
bigquery.SchemaField("client_wrench_id", "STRING", mode="REQUIRED"),
bigquery.SchemaField("expirement_name", "STRING"),
bigquery.SchemaField("processing_datetime", "DATETIME")
]
migration.create_table(name=table_name,
project=migration.client.project,
schema=schema,
dataset=dataset,
partition={'type': 'time'},
clustering_fields=clusters)
return dataset
def down(client):
migration = BigQueryMigration(client)
dataset = migration.dataset('system')
migration.client.query(
f'DELETE FROM {migration.client.project}.system.clients WHERE icentris_client = "bluesun"'
)
return dataset
def down(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name.format(client.env))
table = migration.client.get_table(dataset.table(table_name))
for f in table.schema:
table_field = f.name
if table_field in new_fields:
logging.warning(f'{table_field} already added to {table_name} in {dataset_name} dataset!')
return dataset
def down(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name)
tbls = migration.client.list_tables(dataset)
for item in tbls:
migration.client.delete_table(item)
migration.client.delete_dataset(dataset)
def down(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name)
table = migration.client.get_table(dataset.table(table_name))
if table.schema[-1].name == 'client_paid_rank_level':
logging.warning(
f'Cannot remove rank_level columns from table {table_name}')
return dataset
def down(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name)
table = migration.client.get_table(dataset.table(table_name))
if table.schema[1].mode == 'NULLABLE':
logging.warning(
'leo_eid is already set to NULLABLE. Cannot roll this one back')
return dataset
def down(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name)
table = migration.client.get_table(dataset.table(table_name))
if table.schema[-1].name == 'product_type':
logging.warning(
f'Cannot remove column product_type from table {table_name}')
return dataset
def up(client):
migration = BigQueryMigration(client)
dataset = migration.dataset('system')
sql = f"""
INSERT INTO {migration.client.project}.system.clients (icentris_client, partition_id, wrench_id) VALUES
('bluesun', 4, '5396d067-9e31-4572-951a-a7d1b0a5eaf6')"""
job = migration.client.query(sql)
job_result = job.result()
print(job_result)
return dataset
def down(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name)
table = migration.client.get_table(dataset.table(table_name))
for f in table.schema:
if f.name == 'client_paid_rank' or f.name == 'paid_rank':
if f.mode == 'NULLABLE':
logging.warning(
'paid_rank fields are already set to NULLABLE. Cannot do it again'
)
return dataset
def up(client):
client = BigQueryMigration(client)
dataset = client.dataset(dataset_name)
product_reviews = [
bigquery.SchemaField('client_partition_id', 'INTEGER',
mode='REQUIRED'),
bigquery.SchemaField('client_wrench_id', 'STRING', mode='REQUIRED'),
bigquery.SchemaField('icentris_client', 'STRING', mode='REQUIRED'),
bigquery.SchemaField('site_id', 'INTEGER', mode='REQUIRED'),
bigquery.SchemaField('user_id', 'INTEGER', mode='REQUIRED'),
bigquery.SchemaField('tree_user_id', 'INTEGER'),
bigquery.SchemaField('product_code', 'STRING'),
bigquery.SchemaField('product_name', 'STRING'),
bigquery.SchemaField('product_description', 'STRING'),
bigquery.SchemaField('product_short_description', 'STRING'),
bigquery.SchemaField('product_available_on', 'DATETIME'),
bigquery.SchemaField('product_discontinued_on', 'DATETIME'),
bigquery.SchemaField('product_slug', 'STRING'),
bigquery.SchemaField('product_avg_rating', 'NUMERIC'),
bigquery.SchemaField('product_reviews_count', 'INTEGER'),
bigquery.SchemaField('review_name', 'STRING'),
bigquery.SchemaField('location', 'STRING'),
bigquery.SchemaField('rating', 'INTEGER'),
bigquery.SchemaField('title', 'STRING'),
bigquery.SchemaField('review', 'STRING'),
bigquery.SchemaField('approved', 'BOOLEAN'),
bigquery.SchemaField('created_at', 'DATETIME'),
bigquery.SchemaField('updated_at', 'DATETIME'),
bigquery.SchemaField('ip_address', 'STRING'),
bigquery.SchemaField('show_identifier', 'BOOLEAN'),
]
client.create_table(
name=table_name,
project=client.client.project,
schema=product_reviews,
dataset=dataset,
partition={
'type': 'range',
'field': 'client_partition_id',
'start': 1,
'end': 100,
'interval': 1
},
clustering_fields=['leo_eid:STRING', 'ingestion_timestamp:TIMESTAMP'])
return dataset
def up(client):
client = BigQueryMigration(client)
dataset = client.dataset(dataset_name)
table = client.client.get_table(dataset.table(table_name))
orig_schema = table.schema
new_schema = orig_schema.copy()
new_schema.insert(
1,
bigquery.SchemaField('commission_user_id', 'INTEGER', mode='REQUIRED'))
client.delete_table(table)
client.create_table(
name=table_name,
project=client.client.project,
schema=new_schema,
dataset=dataset,
clustering_fields=['leo_eid:STRING', 'ingestion_timestamp:TIMESTAMP'],
partition={
'type': 'range',
'field': 'client_partition_id',
'start': 1,
'end': 100,
'interval': 1
})
return dataset
def down(client):
client = BigQueryMigration(client)
dataset = client.dataset(dataset_name)
table = client.client.get_table(dataset.table(table_name))
orig_schema = table.schema
new_schema = orig_schema.copy()
if new_schema[1].name == 'commission_user_id':
new_schema.pop(1)
client.delete_table(table)
client.create_table(name=table_name,
project=client.client.project,
schema=new_schema,
dataset=dataset,
partition={
'type': 'range',
'field': 'client_partition_id',
'start': 1,
'end': 100,
'interval': 1
},
clustering_fields=[
'leo_eid:STRING',
'ingestion_timestamp:TIMESTAMP'
])
return dataset
def up(client):
client = BigQueryMigration(client)
dataset = client.create_dataset('staging')
for tbl, schema in schemas.items():
clusters = ['leo_eid:STRING', 'ingestion_timestamp:TIMESTAMP']
client.create_table(name=tbl,
project=client.client.project, schema=schema, dataset=dataset,
partition={
'type': 'range',
'field':
'client_partition_id',
'start': 1,
'end': 100,
'interval': 1},
clustering_fields=clusters)
return dataset
def up(client):
client = BigQueryMigration(client)
dataset = client.dataset(dataset_name)
flat_site_visitors = [
bigquery.SchemaField('client_partition_id', 'INTEGER',
mode='REQUIRED'),
bigquery.SchemaField('client_wrench_id', 'STRING', mode='REQUIRED'),
bigquery.SchemaField('icentris_client', 'STRING', mode='REQUIRED'),
bigquery.SchemaField('site_id', 'INTEGER', mode='REQUIRED'),
bigquery.SchemaField('user_id', 'INTEGER', mode='REQUIRED'),
bigquery.SchemaField('tree_user_id', 'INTEGER'),
bigquery.SchemaField('visitor_id', 'STRING'),
bigquery.SchemaField('last_visit_date', 'DATETIME'),
bigquery.SchemaField('visit_count', 'INTEGER'),
bigquery.SchemaField('ipaddress', 'STRING'),
bigquery.SchemaField('browser_agent', 'STRING'),
bigquery.SchemaField('created_at', 'DATETIME'),
bigquery.SchemaField('site_template_id', 'INTEGER'),
bigquery.SchemaField('active', 'INTEGER'),
bigquery.SchemaField('third_party_tracking_company', 'STRING'),
bigquery.SchemaField('tracking_code', 'STRING'),
bigquery.SchemaField('owner_name', 'STRING'),
bigquery.SchemaField('email', 'STRING'),
bigquery.SchemaField('story', 'STRING'),
bigquery.SchemaField('avatar_file_name', 'STRING')
]
client.create_table(
name=table_name,
project=client.client.project,
schema=flat_site_visitors,
dataset=dataset,
partition={
'type': 'range',
'field': 'client_partition_id',
'start': 1,
'end': 100,
'interval': 1
},
clustering_fields=['leo_eid:STRING', 'ingestion_timestamp:TIMESTAMP'])
return dataset
def up(client):
client = BigQueryMigration(client)
dataset = client.dataset(dataset_name)
# In order to use clusters in BigQuery, the must be inside partitions.
# Even though TIMESTAMP is supported as a partition type,
# a partition can only be done by date not datetime. Furthermore, if timestamp is used
checkpoint_schema = [
bigquery.SchemaField("table", "STRING", mode="REQUIRED"),
bigquery.SchemaField("leo_eid", "STRING",
mode="REQUIRED"), # Partition
bigquery.SchemaField("checkpoint", "TIMESTAMP",
mode="REQUIRED"), # Partition
]
client.create_table(name=table_name,
project=client.client.project,
schema=checkpoint_schema,
dataset=dataset)
return dataset
def up(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name)
table = migration.client.get_table(dataset.table(table_name))
orig_schema = table.schema
new_schema = orig_schema.copy()
if table.schema[-1].name == 'product_type':
logging.warning(
f'product_type already added to {table_name} in {dataset_name} dataset!'
)
return dataset
new_schema.append(bigquery.SchemaField('product_type', 'STRING'))
table.schema = new_schema
migration.client.update_table(table, ['schema'])
return dataset
def up(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name)
table = migration.client.get_table(dataset.table(table_name))
orig_schema = table.schema
new_schema = orig_schema.copy()
for f in table.schema:
table_field = f.name
if table_field in new_fields:
logging.warning(f'{table_field} already added to {table_name} in {dataset_name} dataset!')
new_fields.pop(table_field)
for k, v in new_fields.items():
new_schema.append(bigquery.SchemaField(k, v))
if table.schema != new_schema:
table.schema = new_schema
migration.client.update_table(table, ['schema'])
return dataset
def up(client):
migration = BigQueryMigration(client)
dataset = migration.create_dataset(dataset_name)
parent_dataset = migration.dataset('pyr_bluesun_{}'.format(client.env))
tbls = migration.client.list_tables(parent_dataset)
clusters = migration.default_clustering_fields
clusters.insert(0, 'icentris_client:STRING')
for item in tbls:
tbl = migration.client.get_table(item.reference)
orig = tbl.schema
new = orig.copy()
migration.create_table(name=tbl.table_id,
project=migration.client.project,
schema=new,
dataset=dataset,
partition={'type': 'time'},
clustering_fields=clusters)
return dataset
def up(client):
client = BigQueryMigration(client)
dataset = client.dataset(dataset_name)
schema = [
bigquery.SchemaField("icentris_client", "STRING", mode="REQUIRED"),
bigquery.SchemaField("partition_id", "INTEGER", mode="REQUIRED"),
bigquery.SchemaField("wrench_id", "STRING", mode="REQUIRED")
]
tbl = client.create_table(name=table_name,
project=client.client.project,
schema=schema,
dataset=dataset)
client.client.insert_rows(
client.client.get_table(tbl),
[('monat', 1, '2c889143-9169-436a-b610-48c8fe31bb87'),
('worldventures', 2, 'd7d3e26f-d105-4816-825d-d5858b9cf0d1'),
('naturessunshine', 3, '16bcfb48-153a-4c7d-bb65-19074d9edb17')])
return dataset
def up(client):
migration = BigQueryMigration(client)
name = dataset_name + '_{}'.format(client.env)
with PosixPath('/workspace/bigquery/migrations/bluesun_schema.json').open(
mode='r') as f:
tbls = json.loads(f.read())
dataset = migration.create_dataset(name)
for tbl, raw in tbls.items():
schema = []
for f in raw['fields']:
schema.append(bigquery.SchemaField(f['name'], f['type'],
f['mode']))
migration.create_table(name=tbl,
project=migration.client.project,
schema=schema,
dataset=dataset)
return dataset
def up(client):
migration = BigQueryMigration(client)
dataset = migration.create_dataset(
dataset_name
) # use me if you are creating a new dataset. -- ndg 2/5/20
for tbl, schema in schemas.items():
migration.create_table(
name=tbl,
project=migration.client.project,
schema=schema,
dataset=dataset,
partition={
'type': 'range',
'field': 'client_partition_id',
'start': 1,
'end': 100,
'interval': 1
},
clustering_fields=migration.default_clustering_fields)
return dataset
def up(client):
client = BigQueryMigration(client)
dataset = client.create_dataset('pii')
pii_schema = [
bigquery.SchemaField('client_partition_id', 'INTEGER',
mode='REQUIRED'),
bigquery.SchemaField('client_wrench_id', 'STRING', mode='REQUIRED'),
bigquery.SchemaField("icentris_client", "STRING", mode="REQUIRED"),
bigquery.SchemaField("tree_user_id", "INTEGER"),
bigquery.SchemaField("first_name", "STRING"),
bigquery.SchemaField("last_name", "STRING"),
bigquery.SchemaField("company_name", "STRING"),
bigquery.SchemaField("email", "STRING"),
bigquery.SchemaField("phone", "STRING"),
bigquery.SchemaField("mobile_phone", "STRING"),
bigquery.SchemaField("street", "STRING"),
bigquery.SchemaField("city", "STRING"),
bigquery.SchemaField("state", "STRING"),
bigquery.SchemaField("country", "STRING"),
bigquery.SchemaField("birth_date", "DATE"),
bigquery.SchemaField("gender", "STRING")
]
client.create_table(
name='users',
project=client.client.project,
schema=pii_schema,
dataset=dataset,
partition={
'type': 'range',
'field': 'client_partition_id',
'start': 1,
'end': 100,
'interval': 1
},
clustering_fields=['leo_eid:STRING', 'ingestion_timestamp:TIMESTAMP'])
return dataset
def up(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name)
table = migration.client.get_table(dataset.table(table_name))
orig_schema = table.schema
new_schema = orig_schema.copy()
del new_schema[1]
new_schema.insert(
0, bigquery.SchemaField('dag_id', 'STRING', mode='REQUIRED'))
migration.delete_table(table)
migration.create_table(name=table_name,
project=migration.client.project,
schema=new_schema,
dataset=dataset)
return dataset
def down(client):
migration = BigQueryMigration(client)
dataset = migration.dataset(dataset_name)
table = migration.client.get_table(dataset.table(table_name))
orig_schema = table.schema
new_schema = orig_schema.copy()
if new_schema[0].name == 'dag_id':
new_schema.pop(0)
new_schema.insert(
1, bigquery.SchemaField('leo_eid', 'STRING', mode="NULLABLE"))
migration.delete_table(table)
migration.create_table(name=table_name,
project=migration.client.project,
schema=new_schema,
dataset=dataset)
return dataset
def up(client):
migration = BigQueryMigration(client)
global dataset_name
dataset_name = dataset_name.format(client.env)
wv_ds = migration.create_dataset(dataset_name)
bs_ds = migration.dataset('pyr_bluesun_{}'.format(client.env))
ls = client.list_tables(bs_ds)
for tbl in ls:
tbl_ref = bs_ds.table(tbl.table_id)
tbl = client.get_table(tbl_ref)
migration.create_table(tbl_ref.table_id,
tbl_ref.project,
wv_ds,
schema=tbl.schema)
return wv_ds
def down(client):
client = BigQueryMigration(client)
dataset = client.dataset(dataset_name)
table = client.client.get_table(dataset.table(table_name))
client.delete_table(table)
def down(client):
migration = BigQueryMigration(client)
# required code for deleting a newly created dataset -- ndg 2/5/20
migration.delete_dataset(dataset_name.format(client.env))
