FRDR provides powerful functionality to search for datasets that exist in Canadian research data repositories—both in FRDR itself or in numerous repositories from which FRDR harvests metadata.
A Search text box is available on many of the web pages in FRDR. Search by typing terms into the text box and clicking on the magnifying glass icon. Any search will open a search results page showing relevant datasets as well as available facets with which search results can be refined.
Currently, implemented facets include:
The FRDR search interface is heavily based on discovery code developed and contributed by the University of British Columbia.
There are three ways to specify a Search:
Simply enter one or more search terms into the Search text box. The search results will include datasets with a match on any of the words (logical OR).
Note that accented characters are distinguished from unaccented characters. That is, a search for Quebec will return different results than a search for Québec.
Clicking on the Advanced Search link below the Search text box will open an interface to construct a more advanced query using drop-down menus. A searcher can specify the source (to specify from which repositories to include search results), or specify that terms should be from specific fields such as title, author, description or subject keywords.
Searchers can also specify more complex searches using a subset of the ElasticSearch string query syntax The syntax is powerful, but unforgiving.
An example query might be:
title:(fish AND stock~) AND date:[2016-01-01 TO *]
Exceptions to the ElasticSearch syntax in the FRDR implementation include:
Note that the search query syntax can search any metadata fields in FRDR’s schema, not just the ones available in the Advanced Search interface. E.g.:
This would include any (fully specified) custom metadata fields created for a particular storage group in FRDR. It should be remembered that different repositories expose different metadata fields for harvesting, that there is some disagreement between repositories on how fields should be used, and also that many fields are left unpopulated.
On the search results page each matching dataset discovered is shown with:
Clicking on the name of the dataset takes you the landing page for that dataset. Note that for a dataset that rests in a separate repository, clicking on the dataset name will take you to that other repository where rules and access permissions are no longer controlled by FRDR.
Landing pages vary between repositories, but generally will include:
There may be a significant amount of other metadata available. At the bottom of a FRDR landing page, click on "Show full record" to see all of the metadata that FRDR has about this dataset.
To access the data files themselves for a dataset deposited in FRDR, see the section of the dataset's landing page labelled "Files in Dataset". All the files and folders as submitted can be seen by expanding the folder icon labeled "submitted_data". Some files can be accessed, even previewed, using a web browser. Which file types are viewable depends on your browser's capabilities and settings. Smaller files can be downloaded by your browser. However very large files and datasets with many files or a folder structure will have to be downloaded using Globus Connect.
Getting access to the data files themselves for a dataset in another repository will vary but should be adequately described by that other repository.
The Search results page has a drop down menu to specify the order that searches are displayed. Sorting can be specified by author, title, or date. Note that currently sorting is case-sensitive. Titles (or authors) beginning with a lower case letter (e.g., "dGPS") or with accented characters (e.g., "Évaluation de l'impact du programme...") will appear after "Zooplankton".
By default, Search results displayed are sorted by "relevance"--a weighting scheme that favours matches found in the title firstly, then subject, then description.
FRDR can accommodate additional metadata standards so that more discipline-specific metadata can be saved with data sets and so that searches can be conducted based on the values of those metadata fields. A metadata schema for astronomical observations Common Archive Observation Model (CAOM-2.2) has been added to the Demo version of FRDR as an example, and a data set had been submitted with metadata complying with that standard.
To search specifying such custom metadata, the search query must include the full path to the metadata standard as well as the value(s) for the field(s) being sought. For example the following search query specifies the field value "47" for the single metadata field "sequenceNumber" from the metadata standard "caom2/1.0", which standard has been copied into the Demo instance of FRDR.
That search will find in in Demo a data set from the James Clerk Maxwell Telescope. The metadata fields beyond the minimal Dublin Core standard can be viewed by selecting the "Show full record" button at the bottom of the data set landing page.
Custom metadata fields are not available as menu items on the Advance Search page, nor are they available as search facets. All custom metadata in FRDR are indexed and available for searching.