The dataset viewer is not available for this split.
Rows from parquet row groups are too big to be read: 330.87 MiB (max=286.10 MiB)
Error code:   TooBigContentError

Need help to make the dataset viewer work? Open a discussion for direct support.

The Arabic E-Book Corpus

Alternative Title

مدونة لغوية للكتب العربية الإلكترونية



مدونة الكتب العربية الإلكترونية هي مجموعة بيانات متاحة مجانًا تضم 1,745 كتابًا (81.5 مليون كلمة) نُشرت بواسطة مؤسسة هنداوي بين عامي 2008 و2024. تشمل الكتب أنواعًا مختلفة، بما في ذلك الكتب غير الروائية، الروايات، أدب الأطفال، الشعر، والمسرحيات.


The Arabic E-Book Corpus is a freely available collection of 1,745 books (81.5 million words) published by the Hindawi foundation between 2008 and 2024. The books are of various genres, including non-fiction, novels, children's literature, poetry, and plays.

The dataset was originally published at SND and was added to Hugging Face by Mohammad Fares.



Columns Description

(* = Only applicable to translated works. Empty string in non-translated works.)

  • booknr: An eight-digit unique number assigned by the publisher, corresponding to the file name of the book.

  • cateogy: Category as assigned by the publisher. 213 books in the corpus have two or three labels, in which case they are given as a comma separated list. There are 25 possible values:

    arts                    novels           
    biographies             philosophy       
    business                plays            
    children.stories        poetry           
    detective.fiction       politics         
    economics               psychology       
    environmental.sciences  religions        
    geography               science          
    health                  science.fiction  
    history                 social.sciences  
    linguistics             technology       
    literary.criticism      travel.literature
  • title: The Arabic title of the book.

  • author (string): The Author in Arabic script.

  • pubdate: Publication date in the (yyyy-mm-dd)

  • origtitle*: The title of the original work.

  • origauthor*: The name of the author in the original language from translated books.

  • wc: Word count.

  • origpubdate: Date of the original publication. (yyyy)

  • Arabic name of original language.

  • transdate*: Translation date (yyyy), provided for re-publications of existing translations.

  • translation (TRUE/FALSE): Specifies whether the book is a translation from another language.

  • origpubdate.full (string): Similar as above but with some books specified with

  • origlang: English name of original language.

  • category.main: Similar to category, but for books with two or three category labels, the second and third are omitted, giving one category label for each book.

    • yyyy-yyyy (date range). Represented in orgpubdate by the first year in the rage.
    • unkn: Publication date stated in the work to be unknown. Represented in orgpubdate by NA.
    • mult: Stated in the work to be originally published on multiple (unspecified) dates. Represented in orgpubdate by NA.
  • text: The full text of the book.


Hallberg, A. (2024). The Arabic E-Book Corpus (Version 1) [Data set]. University of Gothenburg. DOI: Available after publication.

Downloads last month
Edit dataset card