0.6.8 Records, Fields, Rows, and Columns

A Structured Dataset Has Repeating Parts

A table becomes easier to reason about when you can name its structure.

Consider:

Song ID Title Artist Genre
S001 Northern Lights Example Band Rock
S002 Side Street Dana Lee Jazz

This dataset contains:

  • records ;
  • fields ;
  • rows ;
  • columns .

These terms are related, but they are not identical.

A Record Represents One Item

A record is a collection of related values describing one item.

In the music table:

Activity diagram showing Row 2 song S001, then Row 3 song S002.
Row2SongS.
  • New
  • Changed
  • Removed

The row is the visible spreadsheet structure.

The record is the logical data item represented by that row.

In a clean table, those ideas line up naturally.

Columns Are the Spreadsheet Representation of Fields

A spreadsheet column commonly represents one field.

For example:

Activity diagram showing Column A Song ID, then Column B Title, then Column C Artist, then Column D Genre.
ColumnASongID.
  • New
  • Changed
  • Removed

Again, the column is the spreadsheet structure.

The field is the meaning assigned to that column.

A Cell Is One Field Value for One Record

A cell sits at the intersection of one row and one column.

Suppose:

B2 = Northern Lights

Column B represents Title.

Row 2 represents song S001.

Therefore the cell represents:

That is a useful way to reason about any structured spreadsheet cell.

CSV Uses the Same Logical Structure Without the Grid

The spreadsheet:

Song ID Title Artist
S001 Northern Lights Example Band

might appear in CSV as:

Song ID,Title,Artist
S001,Northern Lights,Example Band

The grid is gone.

The record-and-field structure remains.

That is why CSV can move structured data between systems.

One Row Should Have One Record Meaning

A table becomes confusing if some rows represent songs and other rows represent artists.

For example:

S001 | Northern Lights | Example Band | Rock
A001 | Example Band    | Milwaukee    | 2018

The first row describes a song.

The second appears to describe an artist.

The columns no longer have one consistent meaning.

If different kinds of records need different fields, separate tables or worksheet tabs may be more appropriate.

One Column Should Have One Field Meaning

Likewise, a column labeled:

Genre

should not contain:

Rock
Jazz
2024
Example Band

if 2024 and Example Band represent different concepts.

A field should have a stable purpose.

Consistency makes later sorting, filtering, and relationships much easier.

Record Structure Supports Questions

Once records and fields are consistent, the dataset can answer useful questions.

For example:

uses the Genre field.

uses the Song ID field.

uses the Artist field for that song record.

Structured questions depend on predictable structure.

Keep the Vocabulary Connected

Use this mental model:

Record

One logical item in the dataset.

Field

One defined piece of information about that item.

Row

The spreadsheet line commonly used to represent one record.

Column

The spreadsheet direction commonly used to represent one field.

Cell

One field value for one record.

Those concepts prepare you to recognize why consistency and unique identifiers matter.