Glossary

This glossary contains key terms and concepts used in OpenRoundup, along with their definitions and explanations. This glossary is designed to help users understand the terminology and concepts that are commonly used in OpenRoundup, and to provide a reference for users who may be new to the system or to data wrangling in general.

Column semantics: The meaning and interpretation of a column in a table, including the type of data it contains, the units of measurement, and the context in which it is used. Understanding column semantics is important for effectively combining tables together and analyzing data, as it helps to ensure that columns are being combined together in a meaningful way and that the resulting data is accurate and useful.
Composite table: A table that is created by combining data from multiple source tables through operations such as stacking and packing. A composite table is the result of a table consolidation process and serves as a unified table schema for analysis.
Delayed table consolidation: A table consolidation approach where data is consolidated into a single composite table after other data wrangling tasks have been performed. This approach emphasizes performing data wrangling tasks on individual tables before combining them into a unified dataset.
Domain constraints: Rules that define the valid values for a column in a relational database. Domain constraints can include data types, value ranges, and allowed values. This mechanisms can help ensure data quality and consistency, but are unavailable in flat data formats used in OpenRoundup.
Dynamic tables: Tables that are periodically updated with new data. Dynamic tables are common in programming languages, such as Python and R, where data is frequently refreshed or appended. In contrast, static tables are published at a single point in time and do not receive updates or changes over time.
Eager table consolidation: A table consolidation approach where all data is consolidated into a single composite table prior to other data wrangling tasks. This approach emphasizes having a complete, unified dataset available for analysis. This approach is one fundamental way that OpenRoundup differ from other research, open-source, and commercial data integration and ETL tools.
Flat tables: Data that is stored in a simple, tabular format, such as ASCII files delimited by special characters (e.g. CSV, TSV) or Excel files. Flat data formats are often used for data exchange and storage, but they lack the structure and metadata of relational databases, such domain constrants, referential integrity, and entity integrity.
Left-hand table: In a Pack operation, the left-hand table is the first table that is selected to be combined. It is positioned on the left side of the Pack operation block in the [Composite Schema View](../interface-overview/composite-schema-view) and the [Schema View](../interface-overview/schema-view).
Pack Operation: A data merging operation that combines tables based on shared columns, similar to a SQL join. Also referred to as packing tables together.
Right-hand table: In a Pack operation, the right-hand table is the second table that is selected to be combined. It is positioned on the right side of the Pack operation block in the [Composite Schema View](../interface-overview/composite-schema-view) and the [Schema View](../interface-overview/schema-view).
Snowball approach: A method of composite table assembly where you start with one table and iteratively combine it with other tables, allowing you to build up a composite table in a step-by-step manner. This approach is one fundamental way that OpenRoundup differ from other research, open-source, and commercial data integration and ETL tools.
Schema: The structure of a table, including the column names, data types, and relationships between columns. A table's schema defines how data is organized and how it can be accessed and analyzed.
Schema Drift: The gradual evolution of a table's schema over time when data is periodically published, fields may be added, removed, renamed, or retyped without notice. Schema drift can silently break downstream pipelines, queries, and reports that were built against an earlier version of the schema. The stack operation is designed to help users identify and resolve schema drift across tables, which is a common issue when consolidating tables that were published at different times.
Schema View: An OpenRoundup view that provides an interactive overview of the focused table or operation's schema. The Schema View enables you to explore the data structurally, understand how tables are being combined by different operations, and make informed decisions about how to combine tables together, while managing complexity. See the Schema View documentation for more details.
Stack Operation: A data merging operation that combines tables by appending rows together, similar to a SQL union. Also referred to as stacking tables together. See the conceptual guide for more details.
Static table: A table that is published at a single point in time and does not receive updates or changes over time. Static tables are often used for one-time analyses or for historical data, and they are a common context for data integration in OpenRoundup. In contrast, dynamic tables are tables that are periodically updated with new data, and they are a common context for data integration in programming languages, such as Python and R.
Table consolidation: The process of combining multiple tables into a single, consolidated table for further analysis and visualization. See the conceptual guide for more details. Within a table consolidation workflow, tables can be combined using either a eager table consolidation approach or a delayed table consolidation approach.
Trimming: Data trimming is the process of removing irrelevant rows and columns from a table. Trimming is an important step in the data wrangling process, as it helps to reduce noise and improve the quality of the data for analysis. OpenRoundup supports limited deletion of rows and columns, but it is often advantageous to perform bulk trimming using external tools before importing data into OpenRoundup.
Wrangling: Data wrangling, also called data munging, data preparation, or data cleaning, is the process of cleaning, transforming, and preparing data for analysis. It involves tasks such as data cleaning, data transformation, data integration, and data enrichment. OpenRoundup is designed to support the exploratory data wrangling process by providing tools for table consolidation and schema management.