The first normalization step consists of 3 conditions:
Let's take a look at a 'books' table using my digital books database:
book title | pubyear | format <=== colunm headings
Book 1|1985,2001|epub <=== WRONG: 2 pubyear values
The Third Book|2021|pdf,epub <=== WRONG: 2 format values
Book the Ninth|2020|azw <=== WRONG: missing unique id
This first attempt seems to make sense, but it contradicts some conditions.
To correct this add each book and their format or pubyear separately, and add a unique ID.
id|book|pubyear|format
1|The Third Book|2021|epub
2|Book 1|1985|epub
3|Book the Ninth|2020|azw
4|Book 1|2001|epub
5|The Third Book|2021|pdf
The way to ensure each line has a unique ID is done during the creation stage using 'integer primary key autoincrement'.
Second Normal Form