Page Comparison

An alternative method to manual entry is parsed entry, which Parsed entry takes existing structured information from the studies (e.g., XML, HTML) and processes it so that the questionnaire can be loaded straight into Archivist. Two types of parsers are used in this process: one to turn the inputted file (e.g., XML, HTML) into tables (note that there are different parsers for the different inputs), and then archivist_insert which loads the tables straight into Archivist. XML is then imported from Archivist or can be downloaded from GitLab (see Figure 1).

This process means the questionnaire needs to be checked and edited, rather than entered from scratch, which speeds up the entry and reduces human error as much less information is entered manually. Details of the parsers and the documents used for each one are included on each individual parser page.

Questionnaires that are parsed into Archivist should be checked to make sure all of the content is included and is in the right order, and that it is nested correctly. Unlike questionnaires which are entered from scratch, we can be reasonably confident that the text within the questionnaire (e.g., the question literals, code list categories, etc.) is correct; it is more important to pay attention to ensuring that all the elements have been entered in the correct location and are nested accurately. Particular attention should be paid to the quirks of each parser. The pages below list each of the parsers, and the specific quirks to look out for for each one.

Note: Although questionnaires are checked before they are worked on, it is possible that you may notice consistent mistakes/omissions when working on a questionnaire. If you notice any consistent errors when checking a questionnaire (e.g., all of the conditions are missing), please raise this with Becky or Hayley before continuing to enter the questionnaire.