Finding an import filter to match your search results can sometimes be tricky. Some of the best databases will have references that are not consistantly formatted. In most cases a lack of consistancy is the reason for a poor import. There are a few things that the user can do to help in the import process.
Saving your search results
1. When saving your search results, the file needs to be in a tagged format. This means that each piece of information should be indentified by some type of prefix or tag. For example, the author information might be identified by a tag such as 'Author' or 'AU: '. This tag is at the beginning of the line. A typical tagged reference might look something like this:
AN: 343708-100300
TI: The five dimensions of social pedagogy within schools.
AU: KYRIACOU, C
AB: This paper discusses social pedagogy in the context of schools, which it defines as taking an integrated view of the whole child in terms of five key dimensions: care and welfare; inclusion; socialisation; academic support; and social education. Considers the implications for professional practice within schools if a social pedagogical approach is adopted and the advantages of such an approach.
DE: SOCIAL PEDAGOGUES, WORKING WITH CHILDREN, SCHOOLS, PUPILS, TEACHERS, PRACTICE, INTERAGENCY RELATIONS, INCLUSION, WELLBEING, SOCIAL SKILLS, ATTITUDES, BEHAVIOUR, PSHE, PERSONAL AND SOCIAL EDUCATION, HEALTH EDUCATION, PASTORAL CARE
SO: Pastoral Care in Education, vol.27, no.2 (Jun). pp101-108. 2009
PT: Journal Article
YR: 2009
UP: 200910
There are hundreds of different formats available. Each database may be different but most should have some type of tagged format available. Also, depending on the portal that you are using to access the database, the format might be different for the same database.
2. If you are given options on the format to save your search results in, look for something that says RIS or RefMan or EndNote or Procite. If you see one of these, then it will be in a format that a filter can easily work with.
3. You often have a choice on the amount of detail you can save for each reference. It is always best to select the one that gives you the most information. Important fields such as the publication type (journal, book, chapter, etc.) are often missing when the option to have less detail is selected.
4. You must be sure that your search result is a text file. This does not mean that it will have a .txt extension but rather it does not contain any formatting. Cutting and pasting search results from a webpage will give you a file full of html formatting. Search results in a Word document will be full of Word formatting. The best way to be sure you have pure text is to cut and paste the content into a text editor such as Notepad and then save it as text. Other useful free text editors are Notepad++ and EditPad Lite. EditPad Lite (http://www.editpadlite.com/) is good because it will allow you to run search and replace functions that span mutliple lines. This is very useful if you need to edit your search results to add missing tags.
Selecting a filter
Once you have uploaded your search result file you can then try to choose an import filter. The filter selection is based on 4 special tags; the start tag, the end tag, the publication type tag and the Source tag.
Each reference needs a unique start tag and end tag. If you look at the example above, the start tag is 'AN: '. In the utility a space character is represented by a dot. Spaces must be considered when choosing a tag. When you choose the start tag from the drop down menu a number of end tags will be presented based on any filters that use the selected start tag.
You should now select an end tag based on the search result file.In this example the end tag could be 'UP: '. End tags are not always consistant so you may need to have a look through the file to see if a particular end tag is used more often than others.
Once an end tag is selected a list of possible filters is then presented. If you click on the 'Select' link next to a filter,the system will try to import the uploaded file. Do not pay too much attention to the names of the import filters. They should only be used as guides. If you get a message saying the number of start tags and end tags are not the same, you may wish to try a different filter or different end tag. In many cases the end tag is not consistant enough throughout the file and you may need to manually add an end tag. This can be done using a text editor that has a search and replace function. If you see possible end tags such as 'EOF: ' or 'EOR: ' or 'ER: ' you may need to edit your file to add these end tags to each reference. Using the multi-line search and replace functions in a text editor such as EditPad Lite (http://www.editpadlite.com/) will allow you to edit a whole file in one command.
You can also narrow down the number of filters by selecting the publication type tag and the source tag. In the above example the publication type tag is 'PT: '. The source tag is 'SO: '. The source tag normally contains a string of information such as journal name, volume, issue, pages and other information.
If you are able to import the file without any error messages you will then want to check to see how well the file was imported. A file could be imported but the information could still be in the wrong fields. To check the import click on the command button labelled 'View imported references'. The references will be displayed on the screen and you can move through each reference by clicking on the Next and Previous buttons.
If the data is not in the correct fields you may wish to try a different filter. If you think that the references are imported well enough you can then export the references.
XML files
Some databases are now making their search results availailable in xml format rather than a normal tagged text file. A file with the XML file extension is an Extensible Markup Language file. An xml file will have start and end tags for each piece of data in the file. These tags will be enclosed in <> brackets. For example the author field might look like this: <author>KYRIACOU, C</author>
We have started adding xml sources to the EXPORT RIS utility. When you upload your file you will see the a new dropdown menu prompting you to 'Select an XML source'. If you have an xml file and your source is listed, it will convert the file to RIS format when you select your source.
Presently we have an xml import filter for WHO International Clinical Trials Registry Platform. If you know of any other sources that you need filters for please let us know.
CSV files
There is now an option to select filters for csv (comma separated variable) files. At the present time the only filter present is for the 'Canadian Health Research Collection' but if there are other databases with csv formats that require a filter please let us know.
Exporting the references
To export the references click on the Export button. This will download an RIS formatted text file of your search results. The RIS format is a fairly consistant format that you should be able to import into most reference management software including EPPI-Reviewer.
Using the example file
If you cut and paste the example reference into a text editor you should be able to upload the file and find a filter to match it.