teradataml-plus

Python Package that extends the functionality of the popular teradataml package through monkey-patching. This is to use field-developed assets more naturally with the existing interface.

Installation

• pip install teradataml-plus

Quickstart

#always import teradata-plus (tdmlplus) first
import tdmlplus

#then import teradataml. It will have all the additional functionality
import teradataml as tdml

# one additional function is for instance to get a correlation matrix straight from the DataFrame, just like in pandas

DF = tdml.DataFrame("some_table")
DF_corr = DF.corr() # not possible withot tdmlplus

History

v0.1.0 (2025-07-25)

• teradataml.DataFrame

  • corr(method="pearson") – correlation matrix like in pandas

• teradataml.random

  • randn(n, mean=0.0, std=1.0) – random normally distributed variables

• teradataml.dba

  • get_amps_count() – get number of AMPs

v0.2.0 (2025-07-30)

• teradataml.DataFrame

  • show_CTE_query() – generate full lineage SQL with CTEs
  • deploy_CTE_view(view_name, replace=False) – create a view from the full CTE SQL
  • easyjoin(other, on, how="left", lsuffix=None, rsuffix=None) – simplified join using common column names with suffix handling

• tdml.dataframe.sql._SQLColumnExpression - aka DataFrameColumn

  • trycast(dtype) – apply TRYCAST SQL expression to a column
  • hashbin(num_bins, salt=None) – compute hash bin from a column with optional salt
  • _power_transform_get_lambda(method="yeo-johnson") – estimate lambda for power transform
  • power_transform(method="yeo-johnson", lambda_val=None) – apply power transform
  • power_fit_transform(method="yeo-johnson") – estimate lambda and transform in one step

• teradataml.random

  • _generate_sql_for_correlated_normals(cov_matrix) – internal SQL generator for correlated normals
  • correlated_normals(df, mean=None, cov=None) – generate synthetic data with correlation structure

• tdml.widgets

  • tab_dfs(dfs) – display multiple DataFrames/tables in widget tabs

• teradataml

  • prettyprint_sql(query) – pretty-print SQL with indentation and keyword formatting

v0.3.0 (2025-08-18)

• teradataml.DataFrame

  • top(n=10, percentage=None) – efficient limiting via Teradata TOP/TOP PERCENT
  • head(n=5, sort_index=False) – overridden to support sort_index; original preserved as _head
  • select_dtypes(include=None, exclude=None) – filter columns by logical dtypes
  • select_tdtypes(include=None, exclude=None) – filter columns by Teradata types
  • histogram(bins=10, exclude_index=True, target_columns=None, groupby_columns=None) – equal-width histograms for numeric columns
  • plot_hist(bins=10, exclude_index=True, target_columns=None, groupby_columns=None, library="plotly", absolute_values=True, percentage_values=False) – plot histograms with Plotly or Seaborn
  • hist(...) – alias for plot_hist
  • categorical_summary(target_columns=None, exclude_index=True, include_percentages=False) – summaries for CHAR/VARCHAR columns
  • column_summary(target_columns=None, exclude_index=True) – general per-column summary
  • fill_RowId(rowid_columnname="row_id") – add sequential row id
  • reset_index(...) – alias for fill_RowId

• tdml.dataframe.sql._SQLColumnExpression - aka DataFrameColumn

  • histogram(bins=10) – column-level histogram with numeric type validation
  • plot_hist(bins=10, library="plotly", absolute_values=True, percentage_values=False, **plotting_args) – column-level plotting wrapper
  • hist(...) – alias for plot_hist
  • map(value_map, keep_original=True, default_else_value=None, output_type=None) – SQL CASE mapping with output type inference