Table reader

Hi there,

In order to use dictionary based tagging I have noticed that I need to use the Table Reader node, which can read only .table file. I have not found yet how to create such a file (.table).

Could anyone help me with this

Hey @teoanton,

what kind of file do you want to read? The Table Reader node is for KNIME data tables only, meaning that it reads only tables written by the Table Writer node.

Cheers,

Julian

1 Like

You can find a video about Table Reader nodes here https://www.knime.com/knime-introductory-course/chapter2/section1/table-reader-node

For dictionary tagging you need a term column and a document column. For Text Processing you can find a documentation here https://www.knime.com/knime-text-processing

Cheers, Iris

Thank you very much. It’s ok now.

One more question: What is the difference between table writer and table creator?

Hi there!

Table Creator node creates data table (you need to manually enter data) which you then use further on in workflow. When you are done and want to save your results in a KNIME table format you will use Table Writer node which will write your data to specified location.

Hope I helped.

Br,
Ivan

1 Like

Yes. Thank you very much

Another set of questions regarding String to Document converter node.

What is the difference between string cell and document cell and what is the use of converting the string cells into document cells?

Many thanks

According to Technical Report The KNIME Text Processing Feature: An Introduction " Textual data which is stored in csv format can be red by the “File Reader” node. All fields of the
documents, such as author, title, abstract, and full text need to separated appropriately and stored in a corresponding column containing String Cells Afterwards the node “Strings to Documents” need to be applied, which creates a DocumentCell for each row. In the dialog of the node it needs to be specified which column contains the data for which field of the document. Thereby the “Strings to Documents” node provides an easy way to convert textual data from csv files, Microsoft Excel files, or also data bases into a list of documents"

What I dont understand is why the conversion must be “in a list of documents”? Why is not enough to remain string cells? Maybe because the following nodes deal only with documents? I guess each document from the list shall have all data included separately (author, title, abstract, and full text)

Thanks

Hey @teoanton,

Documents contain the String in a more structured way. There is meta information like author’s names, document type and publication date stored in the document. The Strings To Document node also tokenizes the input Strings. Meaning, the String will be seperated in Terms which helps to preprocess and transform the text in an easier way.

E.g. the String “KNIME is a data-analytics platform.” will be converted to a document consisting of the terms “KNIME”, “is”, “a”, “data-analytics”, “platform” and “.”. The term structure is necessary for using the tagging nodes. However, the DocumentCell in the data table only shows the String stored in the document. Using the Bag of Words Creator node shows the included terms for each document and also the corresponding tags that have been applied by preceding tagging nodes.

I hope this helps.

Best,
Julian

2 Likes

Thank you for the answer Julian. It was helpful.

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.