dlpy.timeseries.TimeseriesTable¶
-
class
dlpy.timeseries.
TimeseriesTable
(name, timeid=None, groupby_var=None, sequence_opt=None, inputs_target=None, target=None, autoregressive_sequence=None, acc_interval=None, **table_params)¶ Bases: swat.cas.table.CASTable
Table for preprocessing timeseries
It creates an instance of TimeseriesTable by loading from files on the server side, or files on the client side, or in memory CASTable, pandas.DataFrame or pandas.Series. It then performs inplace timeseries formatting, timeseries accumulation, timeseries subsequence generation, and timeseries partitioning to prepare the timeseries into a format that can be followed by subsequent deep learning models.
Parameters: - name : string, optional
Name of the CAS table
- timeid : string, optional
Specifies the column name for the timeid.
Default: None- groupby_var : string or list-of-strings, optional
The groupby variables.
Default: None.- sequence_opt : dict, optional
Dictionary with keys: ‘input_length’, ‘target_length’ and ‘token_size’. It will be created by the prepare_subsequences method.
Default: None- inputs_target : dict, optional
Dictionary with keys: ‘inputs’, ‘target’. It will be created by the prepare_subsequences method.
Default: None
Returns: Attributes: - timeid_type : string
Specifies whether the table uses ‘date’ or ‘datetime’ format
-
__init__
(name, timeid=None, groupby_var=None, sequence_opt=None, inputs_target=None, target=None, autoregressive_sequence=None, acc_interval=None, **table_params)¶ Initialize self. See help(type(self)) for accurate signature.
Methods
__init__(name[, timeid, groupby_var, …]) Initialize self. abs() Return a new CASTable with absolute values of numerics all([axis, bool_only, skipna, level]) Return True for each column with only elements that evaluate to true any([axis, bool_only, skipna, level]) Return True for each column with at least one true element append(other[, ignore_index, …]) Append rows of other to self append_columns(*items, **kwargs) Append variable names to action inputs parameter append_computed_columns(names, code[, inplace]) Append computed columns as specified append_computedvars(*items, **kwargs) Append variable names to computedvars parameter append_computedvarsprogram(*items, **kwargs) Append code to computedvarsprogram parameter append_groupby(*items, **kwargs) Append variable names to groupby parameter append_orderby(*items, **kwargs) Append orderby parameters append_where(*items, **kwargs) Append code to where parameter as_matrix([columns, n]) Represent CASTable as a Numpy array boxplot([column, by]) Make a boxplot from the table data clip([lower, upper, axis]) Clip values at thresholds clip_lower(threshold[, axis]) Clip values at lower threshold clip_upper(threshold[, axis]) Clip values at upper threshold convert_to_sas_time_format(python_time, …) copy([deep, exclude]) Make a copy of the CASTable object corr([method, min_periods]) Compute pairwise correlation of columns count([axis, level, numeric_only]) Return total number of non-missing values in each column create_lags(varname, nlags, byvar) css([casout]) Return the corrected sum of squares of the values of each column cv([casout]) Return the coefficient of variation of the values of each column datastep(code[, casout]) Execute Data step code against the table del_action_params(*names) Delete parameters for specified action names del_param(*keys) Delete parameters del_params(*keys) Delete parameters describe([percentiles, include, exclude, stats]) Get descriptive statistics drop(labels[, axis, level, inplace, errors]) Return a new CASTable object with the specified columns removed drop_duplicates(casout[, subset]) Remove duplicate rows from a CASTable. dropna([axis, how, thresh, subset, inplace]) Drop rows that contain missing values eval(expr[, inplace, kwargs]) Evaluate a CAS table expression exists() Return True if table exists in the server fillna([value, method, axis, inplace, …]) Fill missing values using the specified method find_file_caslib(conn, path) Check whether the specified path is in the caslibs of the current session from_csv(connection, path[, casout]) Create a CASTable from a CSV file from_dict(connection, data[, casout]) Create a CASTable from a dictionary from_items(connection, items[, casout]) Create a CASTable from a (key, value) pairs from_localfile(conn, path[, columns, …]) Create an TimeseriesTable from a file on the client side. from_pandas(conn, pandas_df[, casout]) Create an TimeseriesTable from a pandas DataFrame or Series from_records(connection, data[, casout]) Create a CASTable from records from_serverfile(conn, path[, columns, …]) Create an TimeseriesTable from a file on the server side from_table(tbl[, columns, casout]) Create an TimeseriesTable from a CASTable generate_splitting_code(timeid, start, end, …) get(key[, default]) Get item from object for given key (ex: DataFrame column) get_action_names() Return a list of available CAS actions get_action_params(name, *default) Return parameters for specified action name get_actionset_names() Return a list of available actionsets get_connection() Get the registered connection object get_dtype_counts() Retrieve the frequency of CAS table column data types get_fetch_params() Return options to be used during the table.fetch action get_ftype_counts() Retrieve the frequency of CAS table column data types get_groupby_vars() Return a list of By group variable names get_inputs_param() Return the column names for the inputs= action parameter get_param(key, *default) Return the value of a parameter get_params(*keys) Return the values of one or more parameters get_value(index, col, **kwargs) Retrieve a single scalar value groupby(by[, axis, level, as_index, sort, …]) Specify grouping variables for the table has_groupby_vars() Return True if the table has By group variables configured has_param(*keys) Return True if the specified parameters exist has_params(*keys) Return True if the specified parameters exist head([n, columns, bygroup_as_index, casout]) Retrieve first n rows hist([column, by]) Make a histogram from the table data identify_coltype(col, tbl_colinfo) info([verbose, buf, max_cols, memory_usage, …]) Print summary of CASTable information invoke(_name_, **kwargs) Invoke an action on the registered connection iteritems() Iterate over column names and CASColumn objects iterrows([chunksize]) Iterate over the rows of a CAS table as (index, pandas.Series) pairs itertuples([index, chunksize]) Iterate over rows as tuples kurt([axis, skipna, level, numeric_only, casout]) Return the kurtosis of the values of each column kurtosis([axis, skipna, level, …]) Return the kurtosis of the values of each column lookup(row_labels, col_labels) Retrieve values indicated by row_labels, col_labels positions max([axis, skipna, level, numeric_only, casout]) Return the maximum value of each column mean([axis, skipna, level, numeric_only, casout]) Return the mean value of each column median([axis, skipna, level, numeric_only, …]) Return the median value of each numeric column merge(right[, how, on, left_on, right_on, …]) Merge CASTable objects using a database-style join on a column min([axis, skipna, level, numeric_only, casout]) Return the minimum value of each column mode([axis, numeric_only, max_tie, skipna]) Return the mode of each column next() Return next item in the iteration nlargest(n, columns[, keep, casout]) Return the n largest values ordered by columns nmiss([axis, level, numeric_only, casout]) Return total number of missing values in each column nsmallest(n, columns[, keep, casout]) Return the n smallest values ordered by columns nth(n[, dropna, bygroup_as_index, casout]) Return the nth row nunique([dropna, casout]) Return number of unique elements per column in the CASTable pop(colname) Remove a column from the CASTable and return it prepare_subsequences(seq_len, target[, …]) Prepare the subsequences that will be pass into RNN probt([casout]) Return the p-value of the T-statistics of the values of each column quantile([q, axis, numeric_only, …]) Return values at the given quantile query(expr[, inplace, engine]) Query the table with a boolean expression rename(columns[, errors]) Rename columns of the CASTable. replace([to_replace, value, inplace, limit, …]) Replace values in the data set reset_index([level, drop, inplace, …]) Reset the CASTable index retrieve(_name_, **kwargs) Invoke an action on the registered connection and retrieve results sample([n, frac, replace, weights, …]) Returns a random sample of the table rows select_dtypes([include, exclude, inplace]) Return a subset CASTable including/excluding columns based on data type set_action_params(name, **kwargs) Set parameters for specified action name set_connection(connection) Set the connection to use for action calls set_param(*args, **kwargs) Set paramaters according to key-value pairs set_params(*args, **kwargs) Set paramaters according to key-value pairs skew([axis, skipna, level, numeric_only, casout]) Return the skewness of the values of each column skewness([axis, skipna, level, …]) Return the skewness of the values of each column slice([start, stop, columns, …]) Retrieve the specified rows sort(by[, axis, ascending, inplace, kind, …]) Specify sort parameters for data in a CASTable sort_values(by[, axis, ascending, inplace, …]) Specify sort parameters for data in a CASTable std([axis, skipna, level, ddof, …]) Return the standard deviation of the values of each column stderr([casout]) Return the standard error of the values of each column sum([axis, skipna, level, numeric_only, casout]) Return the sum of the values of each column tail([n, columns, bygroup_as_index, casout]) Retrieve last n rows timeseries_accumlation([acc_interval, …]) Accumulate the TimeseriesTable into regular consecutive intervals timeseries_formatting(timeid, timeseries[, …]) Format the TimeseriesTable timeseries_partition([training_start, …]) Split the dataset into training, validation and testing set to_clipboard(*args, **kwargs) Write the table data to the clipboard to_csv(*args, **kwargs) Write table data to comma-separated values (CSV) to_datastep_params() Create a data step table specification to_dense(*args, **kwargs) Return dense representation of table data to_dict(*args, **kwargs) Convert table data to a Python dictionary to_excel(*args, **kwargs) Write table data to an Excel spreadsheet to_frame([sample_pct, sample_seed, sample, …]) Retrieve entire table as a SASDataFrame to_gbq(*args, **kwargs) Write table data to a Google BigQuery table to_hdf(*args, **kwargs) Write table data to HDF to_html(*args, **kwargs) Render the table data to an HTML table to_input_datastep_params() Create an input data step table specification to_json(*args, **kwargs) Convert the table data to a JSON string to_latex(*args, **kwargs) Render the table data to a LaTeX tabular environment to_msgpack(*args, **kwargs) Write table data to msgpack object to_outtable() Create a copy of the CASTable object with only output table paramaters to_outtable_params() Create a copy of the CASTable parameters using only the output table parameters to_params() Return parameters of CASTable object to_pickle(*args, **kwargs) Pickle (serialize) the table data to_records(*args, **kwargs) Convert table data to record array to_sparse(*args, **kwargs) Convert table data to SparseDataFrame to_sql(*args, **kwargs) Write table records to SQL database to_stata(*args, **kwargs) Write table data to Stata file to_string(*args, **kwargs) Render the table to a console-friendly tabular output to_table() Create a copy of the CASTable object with only input table paramaters to_table_name() Return the name of the table to_table_params() Create a copy of the table parameters containing only input table parameters to_view(*args, **kwargs) Create a view using the current CASTable parameters to_xarray(*args, **kwargs) Represent table data as a numpy.xarray tvalue([casout]) Return the T-statistics for hypothesis testing of the values of each column uss([casout]) Return the uncorrected sum of squares of the values of each column var([axis, skipna, level, ddof, …]) Return the variance of the values of each column with_params(**kwargs) Create copy of table with kwargs inserted as parameters xs(key[, axis, level, copy, drop_level]) Return a cross-section from the CASTable