def create_table(wf_module, table, metadata=None):
if table is None or table.empty:
return StoredObject.__create_empty_table(wf_module, metadata)
else:
hash = hash_table(table)
return StoredObject.__create_table_internal(
wf_module, table, metadata, hash)
def create_table_if_different(wf_module, old_so, table, metadata=None):
if old_so is None:
return StoredObject.create_table(wf_module, table,
metadata=metadata)
hash = hash_table(table)
if hash != old_so.hash:
old_table = old_so.get_table()
if not old_table.equals(table):
return StoredObject.__create_table_internal(wf_module, table,
metadata, hash)
return None
def create_table_if_different(wf_module, old_so, table, metadata=None):
if old_so is None:
return StoredObject.create_table(wf_module,
table,
metadata=metadata)
hash = hash_table(table)
if hash != old_so.hash:
return StoredObject.create_table(wf_module,
table,
metadata=metadata)
else:
return None
def create_table_if_different(cls, wf_module, old_so, table):
hash = hash_table(table)
if (old_so is not None
# Fast: hashes differ, so we don't need to read the table
and hash == old_so.hash
# Slow: compare files. Expensive: reads a file from S3, holds
# both DataFrames in RAM, uses lots of CPU.
and old_so.get_table().equals(table)):
# `table` is identical to what was in `old_so`.
return None
else:
# `table` is new! Yay!
return cls.__create_table_internal(wf_module, table, hash)
def create_table(wf_module, table, metadata=None):
hash = hash_table(table)
return StoredObject.__create_table_internal(wf_module, table,
metadata, hash)
def create_table(cls, wf_module, table):
hash = hash_table(table)
return cls.__create_table_internal(wf_module, table, hash)