def move_column(self, identifier: str, column_id: int, position: int,
datastore: Datastore) -> VizualApiResult:
"""Move a column within a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified column is unknown or the target position invalid.
Parameters
----------
identifier: string
Unique dataset identifier
column_id: int
Unique column identifier
position: int
Target position for the column
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Make sure that position is a valid column index in the new dataset
if position < 0 or position > len(dataset.columns):
raise ValueError('invalid target position \'' + str(position) +
'\'')
# Get index position of column that is being moved
source_idx = dataset.get_index(column_id)
if source_idx is None:
raise ValueError('unknown column identifier \'' + str(column_id) +
'\'')
# No need to do anything if source position equals target position
if source_idx != position:
columns = list(dataset.columns)
columns.insert(position, columns.pop(source_idx))
rows = dataset.fetch_rows()
for row in rows:
row.values.insert(position, row.values.pop(source_idx))
# Store updated dataset to get new identifier
ds = datastore.create_dataset(columns=columns,
rows=rows,
properties={})
return VizualApiResult(ds)
else:
return VizualApiResult(dataset)
def rename_column(self, identifier: str, column_id: int, name: str,
datastore: Datastore) -> VizualApiResult:
"""Rename column in a given dataset.
Raises ValueError if no dataset with given identifier exists, if the
specified column is unknown, or if the given column name is invalid.
Parameters
----------
identifier: string
Unique dataset identifier
column_id: int
Unique column identifier
name: string
New column name
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Raise ValueError if given colum name is invalid
if not is_valid_name(name):
raise ValueError('invalid column name \'' + name + '\'')
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Get the specified column that is to be renamed and set the column name
# to the new name
col_idx = dataset.get_index(column_id)
if col_idx is None:
raise ValueError('unknown column identifier \'' + str(column_id) +
'\'')
# Nothing needs to be changed if name does not differ from column name
if dataset.columns[col_idx].name.lower() != name.lower():
columns = list(dataset.columns)
col = columns[col_idx]
columns[col_idx] = DatasetColumn(identifier=col.identifier,
name=name,
data_type=col.data_type)
# Store updated dataset to get new identifier
ds = datastore.create_dataset(columns=columns,
rows=dataset.fetch_rows(),
properties={})
return VizualApiResult(ds)
else:
return VizualApiResult(dataset)
def move_row(self, identifier: str, row_id: str, position: int,
datastore: Datastore) -> VizualApiResult:
"""Move a row within a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified row or position is not within the range of the dataset.
Parameters
----------
identifier: string
Unique dataset identifier
row_id: int
Global row identifier for deleted row
position: int
Target position for the row
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
command = {"id": "moveRow", "row": row_id, "position": position}
response = mimir.vizualScript(dataset.identifier, command)
return VizualApiResult.from_mimir(response)
def delete_row(self, identifier: str, row_index: str,
datastore: Datastore) -> VizualApiResult:
"""Delete a row in a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified row is not within the range of the dataset.
Parameters
----------
identifier: string
Unique dataset identifier
row_index: int
Row index for deleted row
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Make sure that row refers a valid row in the dataset
if int(row_index) < 0 or int(row_index) >= dataset.row_count:
raise ValueError('invalid row index \'' + str(row_index) + '\'')
# Delete the row at the given index position
rows = dataset.fetch_rows()
del rows[int(row_index)]
# Store updated dataset to get new identifier
ds = datastore.create_dataset(columns=dataset.columns,
rows=rows,
properties={})
return VizualApiResult(ds)
def delete_row(self, identifier: str, row_index: str,
datastore: Datastore) -> VizualApiResult:
"""Delete a row in a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified row is not within the range of the dataset.
Parameters
----------
identifier: string
Unique dataset identifier
row_index: int
Row index for deleted row
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
assert (isinstance(dataset, MimirDatasetHandle))
# Create a view for the modified dataset
col_list = []
for col in dataset.columns:
assert (isinstance(col, MimirDatasetColumn))
col_list.append(col.name_in_rdb)
command = {"id": "deleteRow", "row": int(row_index)}
response = mimir.vizualScript(dataset.identifier, command)
return VizualApiResult.from_mimir(response)
def insert_row(self, identifier: str, position: int,
datastore: Datastore) -> VizualApiResult:
"""Insert row at given position in a dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified row psotion isoutside the dataset bounds.
Parameters
----------
identifier: string
Unique dataset identifier
position: int
Index position at which the row will be inserted
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
command = {"id": "insertRow", "position": position}
response = mimir.vizualScript(dataset.identifier, command)
return VizualApiResult.from_mimir(response)
def delete_column(self, identifier: str, column_id: int,
datastore: Datastore) -> VizualApiResult:
"""Delete a column in a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified column is unknown.
Parameters
----------
identifier: string
Unique dataset identifier
column_id: int
Unique column identifier
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Confirm that the column actually exists and convert the column_identifier to
# a position in the schema (usually ==, but not guaranteed)
col_index = get_index_for_column(dataset, column_id)
command = {"id": "deleteColumn", "column": col_index}
response = mimir.vizualScript(dataset.identifier, command)
return VizualApiResult.from_mimir(response, identifier)
def filter_columns(self, identifier: str, columns: List[int],
names: List[str],
datastore: Datastore) -> VizualApiResult:
"""Dataset projection operator. Returns a copy of the dataset with the
given identifier that contains only those columns listed in columns.
The list of names contains optional new names for the filtered columns.
A value of None in names indicates that the name of the corresponding
column is not changed.
Raises ValueError if no dataset with given identifier exists or if any
of the filter columns are unknown.
Parameters
----------
identifier: string
Unique dataset identifier
columns: list(int)
List of column identifier for columns in the result.
names: list(string)
Optional new names for filtered columns.
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# The schema of the new dataset only contains the columns in the given
# list. Keep track of their index positions to filter values.
schema = list()
val_filter = list()
for i in range(len(columns)):
col_idx = dataset.get_index(columns[i])
if col_idx is None:
raise ValueError('unknown column identifier \'' +
str(columns[i]) + '\'')
col = dataset.columns[col_idx]
if not names[i] is None:
schema.append(
DatasetColumn(identifier=col.identifier,
name=names[i],
data_type=col.data_type))
else:
schema.append(col)
val_filter.append(col_idx)
# Create a list of projected rows
rows = list()
for row in dataset.fetch_rows():
values = list()
for v_idx in val_filter:
values.append(row.values[v_idx])
rows.append(DatasetRow(identifier=row.identifier, values=values))
# Store updated dataset to get new identifier
ds = datastore.create_dataset(columns=schema, rows=rows, properties={})
return VizualApiResult(ds)
def update_cell(self, identifier, column_id, row_id, value, datastore):
"""Update a cell in a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified cell is outside of the current dataset ranges.
Parameters
----------
identifier : string
Unique dataset identifier
column_id: int
Unique column identifier for updated cell
row_id: int
Unique row identifier
value: string
New cell value
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Get the index of the specified cell column
col_index = get_index_for_column(dataset, column_id)
# Raise exception if row id is not valid
# Create a view for the modified dataset
col_list = []
for i in range(len(dataset.columns)):
col = dataset.columns[i]
if i == col_index:
try:
val_stmt = col.to_sql_value(value)
col_sql = val_stmt + ' ELSE ' + col.name_in_rdb + ' END '
except ValueError:
col_sql = '\'' + str(
value
) + '\' ELSE CAST({{input}}.' + col.name_in_rdb + ' AS varchar) END '
rid_sql = MIMIR_ROWID_COL.to_sql_value(row_id)
stmt = 'CASE WHEN ' + ROW_ID + ' = ' + rid_sql + ' THEN '
stmt += col_sql
stmt += 'AS ' + col.name_in_rdb
col_list.append(stmt)
else:
col_list.append(col.name_in_rdb)
sql = 'SELECT ' + ','.join(
col_list) + ' FROM ' + dataset.table_name + ';'
view_name, dependencies = mimir.createView(dataset.table_name, sql)
# Store updated dataset information with new identifier
ds = datastore.register_dataset(table_name=view_name,
columns=dataset.columns,
row_counter=dataset.row_counter,
annotations=dataset.annotations)
return VizualApiResult(ds)
def move_column(self, identifier, column_id, position, datastore):
"""Move a column within a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified column is unknown or the target position invalid.
Parameters
----------
identifier: string
Unique dataset identifier
column_id: int
Unique column identifier
position: int
Target position for the column
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Make sure that position is a valid column index in the new dataset
if position < 0 or position > len(dataset.columns):
raise ValueError('invalid target position \'' + str(position) +
'\'')
# Get index position of column that is being moved
source_idx = get_index_for_column(dataset, column_id)
# No need to do anything if source position equals target position
if source_idx != position:
# There are no changes to the underlying database. We only need to
# change the column information in the dataset schema.
schema = list(dataset.columns)
schema.insert(position, schema.pop(source_idx))
# Store updated dataset to get new identifier
ds = datastore.register_dataset(table_name=dataset.table_name,
columns=schema,
row_counter=dataset.row_counter,
annotations=dataset.annotations)
return VizualApiResult(ds)
else:
return VizualApiResult(dataset)
def sort_dataset(self, identifier: str, columns: List[int],
reversed: List[bool],
datastore: Datastore) -> VizualApiResult:
"""Sort the dataset with the given identifier according to the order by
statement. The order by statement is a pair of lists. The first list
contains the identifier of columns to sort on. The second list contains
boolean flags, one for each entry in columns, indicating whether sort
order is revered for the corresponding column or not.
Returns the number of rows in the dataset and the identifier of the
sorted dataset.
Raises ValueError if no dataset with given identifier exists or if any
of the columns in the order by clause are unknown.
Parameters
----------
identifier: string
Unique dataset identifier
columns: list(int)
List of column identifier for sort columns.
reversed: list(bool)
Flags indicating whether the sort order of the corresponding column
is reveresed.
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Fetch the full set of rows
rows = dataset.fetch_rows()
# Sort multiple times, ones for each of the sort columns (in reverse
# order of appearance in the order by clause)
for i in range(len(columns)):
l_idx = len(columns) - (i + 1)
col_id = columns[l_idx]
col_idx = dataset.get_index(col_id)
# print("SORT: {}".format(col_idx))
# print("\n".join(", ".join("'{}':{}".format(v, type(v)) for v in row.values) for row in rows))
reverse = reversed[l_idx]
if col_idx is None:
raise ValueError('unknown column identifier \'' + str(col_id) +
'\'')
else:
rows.sort(key=lambda row: row.values[cast(int, col_idx)],
reverse=reverse)
# Store updated dataset to get new identifier
ds = datastore.create_dataset(columns=dataset.columns,
rows=rows,
properties={})
return VizualApiResult(ds)
def insert_column(self, identifier, position, name, datastore):
"""Insert column with given name at given position in dataset.
Raises ValueError if no dataset with given identifier exists, if the
specified column position is outside of the current schema bounds, or if
the column name is invalid.
Parameters
----------
identifier: string
Unique dataset identifier
position: int
Index position at which the column will be inserted
name: string, optional
New column name
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Raise ValueError if given colum name is invalid
if not is_valid_name(name):
raise ValueError('invalid column name \'' + str(name) + '\'')
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Make sure that position is a valid column index in the new dataset
if position < 0 or position > len(dataset.columns):
raise ValueError('invalid column index \'' + str(position) + '\'')
# Get identifier for new column
col_id = dataset.max_column_id() + 1
# Insert new column into schema
schema = list(dataset.columns)
new_column = MimirDatasetColumn(col_id, name, name)
schema.insert(position, new_column)
# Create a view for the modified schema
col_list = []
for col in schema:
if col.identifier == new_column.identifier:
# Note: By no (April 2018) this requires Mimir to run with the
# XNULL option. Otherwise, in some scenarios setting the all
# values in the new column to NULL may cause an exception.
col_list.append(" CAST('' AS int) AS " + col.name_in_rdb)
else:
col_list.append(col.name_in_rdb)
sql = 'SELECT ' + ','.join(
col_list) + ' FROM ' + dataset.table_name + ';'
view_name, dependencies = mimir.createView(dataset.table_name, sql)
# Store updated dataset information with new identifier
ds = datastore.register_dataset(table_name=view_name,
columns=schema,
row_counter=dataset.row_counter,
annotations=dataset.annotations)
return VizualApiResult(ds)
def filter_columns(self, identifier: str, columns: List[int],
names: List[str],
datastore: Datastore) -> VizualApiResult:
"""Dataset projection operator. Returns a copy of the dataset with the
given identifier that contains only those columns listed in columns.
The list of names contains optional new names for the filtered columns.
A value of None in names indicates that the name of the corresponding
column is not changed.
Raises ValueError if no dataset with given identifier exists or if any
of the filter columns are unknown.
Parameters
----------
identifier: string
Unique dataset identifier
columns: list(int)
List of column identifier for columns in the result.
names: list(string)
Optional new names for filtered columns.
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# The schema of the new dataset only contains the columns in the given
# list. A column might need to be renamed.
schema = list()
column_mapping = list()
col_list = []
for i in range(len(columns)):
col_idx = get_index_for_column(dataset, columns[i])
col = dataset.columns[col_idx]
if not names[i] is None:
if not is_valid_name(names[i]):
raise ValueError('invalid column name \'' + str(names[i]) +
'\'')
schema.append(
MimirDatasetColumn(identifier=col.identifier,
name_in_dataset=names[i],
name_in_rdb=names[i]))
else:
schema.append(col)
column_mapping.append({
"columns_column": col_idx,
"columns_name": schema[-1].name
})
col_list.append(col.name_in_rdb)
command = {"id": "projection", "columns": column_mapping}
response = mimir.vizualScript(dataset.identifier, command)
return VizualApiResult.from_mimir(response)
def move_row(self, identifier: str, row_id: str, position: int,
datastore: Datastore):
"""Move a row within a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified row or position is not within the range of the dataset.
Parameters
----------
identifier: string
Unique dataset identifier
row_index: int
Row index for deleted row
position: int
Target position for the row
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Make sure that row is within dataset bounds
if int(row_id) < 0 or int(row_id) >= dataset.row_count:
raise ValueError('invalid source row \'' + str(row_id) + '\'')
# Make sure that position is a valid row index in the new dataset
if position < 0 or position > dataset.row_count:
raise ValueError('invalid target position \'' + str(position) +
'\'')
# No need to do anything if source position equals target position
if row_id != position:
rows = dataset.fetch_rows()
rows.insert(position, rows.pop(int(row_id)))
# Store updated dataset to get new identifier
ds = datastore.create_dataset(columns=dataset.columns,
rows=rows,
properties={})
return VizualApiResult(ds)
else:
return VizualApiResult(dataset)
def move_column(self, identifier: str, column_id: int, position: int,
datastore: Datastore) -> VizualApiResult:
"""Move a column within a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified column is unknown or the target position invalid.
Parameters
----------
identifier: string
Unique dataset identifier
column_id: int
Unique column identifier
position: int
Target position for the column
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Make sure that position is a valid column index in the new dataset
if position < 0 or position > len(dataset.columns):
raise ValueError('invalid target position \'' + str(position) +
'\'')
# Get index position of column that is being moved
source_idx = get_index_for_column(dataset, column_id)
# No need to do anything if source position equals target position
if source_idx != position:
# Keep the mimir-side schema aligned with the vizier-side schema
command = {
"id": "moveColumn",
"column": source_idx,
"position": position
}
response = mimir.vizualScript(dataset.identifier, command)
return VizualApiResult.from_mimir(response)
else:
return VizualApiResult(dataset)
def filter_columns(self, identifier, columns, names, datastore):
"""Dataset projection operator. Returns a copy of the dataset with the
given identifier that contains only those columns listed in columns.
The list of names contains optional new names for the filtered columns.
A value of None in names indicates that the name of the corresponding
column is not changed.
Raises ValueError if no dataset with given identifier exists or if any
of the filter columns are unknown.
Parameters
----------
identifier: string
Unique dataset identifier
columns: list(int)
List of column identifier for columns in the result.
names: list(string)
Optional new names for filtered columns.
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# The schema of the new dataset only contains the columns in the given
# list. A column might need to be renamed.
schema = list()
col_list = []
for i in range(len(columns)):
col_idx = get_index_for_column(dataset, columns[i])
col = dataset.columns[col_idx]
if not names[i] is None:
schema.append(
MimirDatasetColumn(identifier=col.identifier,
name_in_dataset=names[i],
name_in_rdb=col.name_in_rdb))
else:
schema.append(col)
col_list.append(col.name_in_rdb)
sql = 'SELECT ' + ','.join(
col_list) + ' FROM ' + dataset.table_name + ';'
view_name, dependencies = mimir.createView(dataset.table_name, sql)
# Store updated dataset information with new identifier
ds = datastore.register_dataset(table_name=view_name,
columns=schema,
row_counter=dataset.row_counter,
annotations=dataset.annotations.filter(
columns=columns,
rows=dataset.row_ids))
return VizualApiResult(ds)
def move_row(self, identifier, row_index, position, datastore):
"""Move a row within a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified row or position is not within the range of the dataset.
Parameters
----------
identifier: string
Unique dataset identifier
row_index: int
Row index for deleted row
position: int
Target position for the row
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Make sure that row is within dataset bounds
if row_index < 0 or row_index >= dataset.row_count:
raise ValueError('invalid source row \'' + str(row_index) + '\'')
# Make sure that position is a valid row index in the new dataset
if position < 0 or position > dataset.row_count:
raise ValueError('invalid target position \'' + str(position) +
'\'')
# No need to do anything if source position equals target position
# Store updated dataset to get new identifier
ds = datastore.register_dataset(table_name=dataset.table_name,
columns=dataset.columns,
row_counter=dataset.row_counter,
annotations=dataset.annotations)
return VizualApiResult(ds)
else:
return VizualApiResult(dataset)
def materialize_dataset(self, identifier: str,
datastore: Datastore) -> VizualApiResult:
"""Create a materialized snapshot of the dataset for faster
execution.
This is a no-op for the FS Backend
"""
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
return VizualApiResult(dataset)
def update_cell(self, identifier: str, column_id: int, row_id: str,
value: str, datastore: Datastore) -> VizualApiResult:
"""Update a cell in a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified cell is outside of the current dataset ranges.
Parameters
----------
identifier : string
Unique dataset identifier
column_id: int
Unique column identifier for updated cell
row_id: int
Unique row identifier
value: string
New cell value
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Get column index forst in case it raises an exception
col_idx = dataset.get_index(column_id)
if col_idx is None:
raise ValueError('unknown column identifier \'' + str(column_id) +
'\'')
# Update the specified cell in the given data array
rows = dataset.fetch_rows()
row_index = -1
for i in range(len(rows)):
if int(rows[i].identifier) == int(row_id):
row_index = i
break
# Make sure that row refers a valid row in the dataset
if row_index < 0:
raise ValueError('invalid row identifier \'' + str(row_id) + '\'')
r = rows[row_index]
values = list(r.values)
values[col_idx] = value
rows[row_index] = DatasetRow(identifier=r.identifier, values=values)
# Store updated dataset to get new identifier
ds = datastore.create_dataset(columns=dataset.columns,
rows=rows,
properties={})
return VizualApiResult(ds)
def sort_dataset(self, identifier: str, columns: List[int],
reversed: List[bool],
datastore: Datastore) -> VizualApiResult:
"""Sort the dataset with the given identifier according to the order by
statement. The order by statement is a pair of lists. The first list
contains the identifier of columns to sort on. The second list contains
boolean flags, one for each entry in columns, indicating whether sort
order is revered for the corresponding column or not.
Returns the number of rows in the dataset and the identifier of the
sorted dataset.
Raises ValueError if no dataset with given identifier exists or if any
of the columns in the order by clause are unknown.
Parameters
----------
identifier: string
Unique dataset identifier
columns: list(int)
List of column identifier for sort columns.
reversed: list(bool)
Flags indicating whether the sort order of the corresponding column
is reveresed.
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
assert (isinstance(dataset, MimirDatasetHandle))
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Create order by clause based on columns and reversed flags
order_by_clause = list()
for i in range(len(columns)):
col_id = columns[i]
stmt = cast(MimirDatasetColumn,
dataset.column_by_id(col_id)).name_in_rdb
if reversed[i]:
stmt += ' DESC'
order_by_clause.append(stmt)
sql = 'SELECT * FROM ' + dataset.identifier + ' ORDER BY '
sql += ','.join(order_by_clause)
view_name, dependencies, schema, properties, functionDeps = mimir.createView(
datasets={dataset.identifier: dataset.identifier}, query=sql)
ds = MimirDatasetHandle.from_mimir_result(view_name, schema,
properties)
return VizualApiResult(ds)
def import_dataset(self, datastore: Datastore, project_id: str,
dataset_id: str) -> VizualApiResult:
from vizier.api.webservice.server import api
# Mimir doesn't actually need to use the project ID (yet), but let's check the
# URL for safety anyway
project = api.projects.projects.get_project(project_id)
if project is None:
raise Exception("No Such Project: {}".format(project_id))
# Get the actual dataset
dataset = datastore.get_dataset(dataset_id)
if dataset is None:
raise Exception("No Such Dataset: {}".format(dataset_id))
return VizualApiResult(dataset, {})
def sort_dataset(self, identifier, columns, reversed, datastore):
"""Sort the dataset with the given identifier according to the order by
statement. The order by statement is a pair of lists. The first list
contains the identifier of columns to sort on. The second list contains
boolean flags, one for each entry in columns, indicating whether sort
order is revered for the corresponding column or not.
Returns the number of rows in the dataset and the identifier of the
sorted dataset.
Raises ValueError if no dataset with given identifier exists or if any
of the columns in the order by clause are unknown.
Parameters
----------
identifier: string
Unique dataset identifier
columns: list(int)
List of column identifier for sort columns.
reversed: list(bool)
Flags indicating whether the sort order of the corresponding column
is reveresed.
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Create order by clause based on columns and reversed flags
order_by_clause = list()
for i in range(len(columns)):
col_id = columns[i]
stmt = dataset.column_by_id(col_id).name_in_rdb
if reversed[i]:
stmt += ' DESC'
order_by_clause.append(stmt)
sql = 'SELECT * FROM {{input}} ORDER BY '
sql += ','.join(order_by_clause) + ';'
view_name, dependencies = mimir.createView(dataset.table_name, sql)
# Register new dataset with only a modified list of row identifier
ds = datastore.register_dataset(table_name=view_name,
columns=dataset.columns,
annotations=dataset.annotations)
return VizualApiResult(ds)
def insert_column(self, identifier: str, position: int, name: str,
datastore: Datastore) -> VizualApiResult:
"""Insert column with given name at given position in dataset.
Raises ValueError if no dataset with given identifier exists, if the
specified column position is outside of the current schema bounds, or if
the column name is invalid.
Parameters
----------
identifier: string
Unique dataset identifier
position: int
Index position at which the column will be inserted
name: string
New column name
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Raise ValueError if given colum name is invalid
if not name is None and not is_valid_name(name):
raise ValueError('invalid column name \'' + name + '\'')
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Make sure that position is a valid column index in the new dataset
if position < 0 or position > len(dataset.columns):
raise ValueError('invalid column index \'' + str(position) + '\'')
# Insert new column into dataset
columns = list(dataset.columns)
rows = dataset.fetch_rows()
columns.insert(
position,
DatasetColumn(identifier=dataset.max_column_id() + 1,
name=name if not name is None else ''))
# Add a null value to each row for the new column
for row in rows:
row.values.insert(position, None)
# Store updated dataset to get new identifier
ds = datastore.create_dataset(columns=columns,
rows=rows,
properties={})
return VizualApiResult(ds)
def materialize_dataset(self, identifier: str,
datastore: Datastore) -> VizualApiResult:
"""Create a materialized snapshot of the dataset for faster
execution."""
input_dataset = datastore.get_dataset(identifier)
if input_dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
cast(MimirDatasetHandle, input_dataset)
response = mimir.materialize(input_dataset.identifier)
output_ds = MimirDatasetHandle(
identifier=response["name"],
columns=cast(List[MimirDatasetColumn], input_dataset.columns),
properties=input_dataset.get_properties(),
name=input_dataset.name
if input_dataset.name is not None else "untitled dataset")
return VizualApiResult(output_ds)
def insert_row(self, identifier, position, datastore):
"""Insert row at given position in a dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified row psotion isoutside the dataset bounds.
Parameters
----------
identifier: string
Unique dataset identifier
position: int
Index position at which the row will be inserted
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Make sure that position is a valid row index in the new dataset
if position < 0 or position > len(dataset.row_ids):
raise ValueError('invalid row index \'' + str(position) + '\'')
# Get unique id for new row
dataset.row_counter += 1
# Create a view for the modified schema
col_list = []
for col in dataset.columns:
col_list.append(col.name_in_rdb)
sql = 'SELECT ' + ','.join(col_list) + ' FROM ' + dataset.table_name
mimirSchema = mimir.getSchema(sql)
union_list = []
for col in mimirSchema[1:]:
union_list.append('CAST(NULL AS ' + col['baseType'] + ') AS ' +
col['name'])
sql = '(' + sql + ') UNION ALL (SELECT ' + ','.join(union_list) + ');'
view_name, dependencies = mimir.createView(dataset.table_name, sql)
# Store updated dataset information with new identifier
ds = datastore.register_dataset(table_name=view_name,
columns=dataset.columns,
row_counter=dataset.row_counter,
annotations=dataset.annotations)
return VizualApiResult(ds)
def unload_dataset(self,
dataset: DatasetDescriptor,
datastore: Datastore,
filestore: Filestore,
unload_format: str = 'csv',
options: List[Dict[str, Any]] = [],
resources: Dict[str, Any] = None):
"""Export (or unload) a dataset to a given file format.
The resources refer to any resoures (e.g., file identifier) that have
been generated by a previous execution of the respective task. This
allows to associate an identifier with a downloaded file to avoid future
downloads (unless the reload flag is True).
Parameters
----------
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
filestore: vizier.filestore.Filestore
Filestore to retrieve uploaded datasets
unload_format: string, optional
Format identifier
options: list, optional
Additional options for Mimirs load command
resources: dict, optional
Dictionary of additional resources (i.e., key,value pairs) that were
generated during a previous execution of the associated module
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
f_handles = None
result_resources = dict()
assert (isinstance(datastore, MimirDatastore))
assert (isinstance(filestore, FileSystemFilestore))
if dataset is not None:
f_handles = datastore.unload_dataset(
filepath=filestore.get_file_dir(get_unique_identifier()),
dataset_name=dataset.identifier,
format=unload_format,
options=options)
result_resources[base.RESOURCE_FILEID] = f_handles
return VizualApiResult(dataset=dataset, resources=result_resources)
def empty_dataset(
self,
datastore: Datastore,
filestore: Filestore,
initial_columns: List[Tuple[str, str]] = [("''", "unnamed_column")]
) -> VizualApiResult:
"""Create (or load) a new dataset from a given file or Uri. It is
guaranteed that either the file identifier or the url are not None but
one of them will be None. The user name and password may only be given
if an url is given.
The resources refer to any resoures (e.g., file identifier) that have
been generated by a previous execution of the respective task. This
allows to associate an identifier with a downloaded file to avoid future
downloads (unless the reload flag is True).
Parameters
----------
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
filestore: vizier.filestore.Filestore
Filestore to retrieve uploaded datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
assert (isinstance(datastore, MimirDatastore))
ds = datastore.create_dataset(
columns=[
MimirDatasetColumn(identifier=id,
name_in_dataset=col,
data_type="varchar")
for id, (default, col) in enumerate(initial_columns)
],
rows=[
DatasetRow(
identifier=str(id),
values=[default for default, col in initial_columns])
for id in range(1, 2)
],
human_readable_name="Empty Table",
)
return VizualApiResult(dataset=ds)
def delete_column(self, identifier, column_id, datastore):
"""Delete a column in a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified column is unknown.
Parameters
----------
identifier: string
Unique dataset identifier
column_id: int
Unique column identifier
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Get the index of the specified column that is to be deleted.
col_index = dataset.get_index(column_id)
if col_index is None:
raise ValueError('unknown column identifier \'' + str(column_id) +
'\'')
# Delete column from schema. Keep track of the column name for the
# result output.
columns = list(dataset.columns)
name = columns[col_index].name
del columns[col_index]
# Delete all value for the deleted column
rows = dataset.fetch_rows()
for row in rows:
del row.values[col_index]
# Store updated dataset to get new identifier
ds = datastore.create_dataset(
columns=columns,
rows=rows,
annotations=dataset.annotations.filter(
columns=[c.identifier for c in columns]))
return VizualApiResult(ds)
def insert_column(self, identifier: str, position: int, name: str,
datastore: Datastore) -> VizualApiResult:
"""Insert column with given name at given position in dataset.
Raises ValueError if no dataset with given identifier exists, if the
specified column position is outside of the current schema bounds, or if
the column name is invalid.
Parameters
----------
identifier: string
Unique dataset identifier
position: int
Index position at which the column will be inserted
name: string, optional
New column name
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Raise ValueError if given colum name is invalid
if not is_valid_name(name):
raise ValueError('invalid column name \'' + str(name) + '\'')
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Make sure that position is a valid column index in the new dataset
if position < 0 or position > len(dataset.columns):
raise ValueError('invalid column index \'' + str(position) + '\'')
# Get identifier for new column
col_id = dataset.max_column_id() + 1
# Insert new column into schema
schema = list(dataset.columns)
new_column = MimirDatasetColumn(col_id, name, name)
schema.insert(position, new_column)
command = {"id": "insertColumn", "name": name, "position": position}
response = mimir.vizualScript(dataset.identifier, command)
return VizualApiResult.from_mimir(response)
def delete_column(self, identifier, column_id, datastore):
"""Delete a column in a given dataset.
Raises ValueError if no dataset with given identifier exists or if the
specified column is unknown.
Parameters
----------
identifier: string
Unique dataset identifier
column_id: int
Unique column identifier
datastore : vizier.datastore.fs.base.FileSystemDatastore
Datastore to retireve and update datasets
Returns
-------
vizier.engine.packages.vizual.api.VizualApiResult
"""
# Get dataset. Raise exception if dataset is unknown
dataset = datastore.get_dataset(identifier)
if dataset is None:
raise ValueError('unknown dataset \'' + identifier + '\'')
# Get the index of the specified column that is to be deleted.
col_index = get_index_for_column(dataset, column_id)
# Delete column from schema
schema = list(dataset.columns)
del schema[col_index]
# Create a view for the modified schema
col_list = []
for col in schema:
col_list.append(col.name_in_rdb)
sql = 'SELECT ' + ','.join(
col_list) + ' FROM ' + dataset.table_name + ';'
view_name, dependencies = mimir.createView(dataset.table_name, sql)
# Store updated dataset information with new identifier
ds = datastore.register_dataset(table_name=view_name,
columns=schema,
row_counter=dataset.row_counter,
annotations=dataset.annotations)
return VizualApiResult(ds)
