Advanced Keyword Search

Like the keyword search the advanced keyword search allows you to search over multiple fields for keywords that appear in each field. For each field, each word you enter is considered a single keyword. Where the advanced search differs, however, is that it allows you to specify the relationship between the keywords within each field. The keyword search only needs one or more of your keywords to appear in each field; using the advanced keyword search you can specify that more than one of your keywords must appear in each field.

To specify the relationship between keywords you can use the special characters & and ,. The & character between two keywords indicates that you require both of those keywords to be present in the field; the , character between two keywords indicates that you require either keyword to be present. For example, if you enter India&Pakistan in a search field, you will receive a list of members for whom that field contains both India and Pakistan - both must be present; if, however, you enter India,Pakistan you will receive a list of members for whom that field contains either India or Pakistan or both. (This latter case works much the same as a search using the normal keyword strategy.)

Like the keyword search, if you enter criteria in more than one field, a member's record must match your criteria in both fields to be included in your results, so searching for Rohan in the First Name field and India&Pakistan in the Interests field will provide you with a list of members whose first name contains Rohan and whose interests contain both India and Pakistan.

More complex searches can be constructed using the advanced keyword search by specifying multiple keywords using multiple instances of one or both of the special characters. For example, if you wish to retrieve a list of members whose interests include both India and Pakistan, or whose interests contain Postcolonialism, you can enter India&Pakistan,Postcolonialism. The results list will include members whose interests include both India and Pakistan, members whose interests include Postcolonialism and members whose interests include all three. Importantly, this search will not include members whose interests contain India but not Pakistan, nor will it include members whose interests include Pakistan but not India. In a similar way, you can search for India&Pakistan&Postcolonialism to search for members whose interests include all three of India, Pakistan and Poscolonialsm.

If you wish you may include spacing around your keywords. For relatively simple instances of advanced keyword searches this has limited utility but if you are constructing complex searches involving multiple keywords, spacing out your search criteria can help readability. Searching for India & Pakistan & Postcolonialism is the same as searching for India&Pakistan&Postcolonialism.

The advanced keyword search does not perform any analysis of your results and it is not currently possible to order the results by how relevant they are. For most other search strategies this is not particularly important, but the complexity of the searches that can be constructed using the advanced keyword strategy makes ranking by relevance more important. This is something we are working on and hopefully it will be available soon.

Searches carried out using the advanced keyword search strategy are not case sensitive. This means that the keywords india, INDIA, India and INdiA are all equivalent and will produce the same results.

Wildcards

Wildcards are special characters you can type in your search keywords to perform special functions. Two wildcards are supported with the advanced keyword search:

• the * character will match any number of any character, including no characters at all
• the ? character will match any single character

Using wildcards enables basic pattern matching to be carried out in your searches. For example, searching for post*ism means search for any piece of text that starts with "post", then contains any combination of any number of characters and ends with "ism". This would match postcolonialism, poststructuralism and postmodernism and so your search would include results where any of those three terms is present. This is because the "*" in the keyword can match any number of characters, so it matches colonial in postcolonialism, structural in poststructuralism and modern in postmodernism. Similarly, searching for cl?p translates as match any piece of text that starts with "cl", then has any one character, then ends with "p". This would match clap and clip because the "?" can match the a in clap and the i in clip; however, it would not match clasp as the "?" cannot match the pair of characters as.

The ? wildcard can be useful in cases where you are unsure of the spelling of a word (for example where the word is prone to mis-spelling or is spelled differrently in different language locales). For example, a speaker of British English may be interested in modernisation whereas a speaker of American English would be interested in modernization. The ? wildcard can help accommodate this difference: both words would match a search for moderni?ation.

Use of multiple wildcards is permitted as is mixing different wildcards in a single field or even in a single keyword. Keywords such as cl??t (match any piece of text that starts with "cl" followed by one of any character, followed by one of any character, followed by "t") are valid to match clout and cleat. (Note that the two "?" wildcards may match different characters.) It is possible to use multiple "*" wildcards as well, although this is less common and generally less useful as it tends to result in too many matches.

There are some pitfalls to be aware of when using the advanced keyword search, particularly with wildcards. The first is that it does not just search for whole-word matches, it searches for part-word matches. This is something of a double-edged sword because it means you can use the keyword india and you will receive results containing india or indian, which is likely to be useful more often than not; however, it also means that searching for rite will match rite, writer, trite and so on, which is less likely to be of use.

A slightly more subtle but potentially more invasive issue is that whitespace and punctuation are valid characters when matching wildcards. In the example above, as well as providing the results mentioned, searching for post*ism will also match the post was late, which has become something of a truism because that phrase contains "post" followed by other characters, followed by "ism".

Finally, it is fine to combine wildcards and the special characters & and , in the same field. In fact, using the & special character makes employing multiple * wildcards somewhat more useful for complex searches.

International alphabets

The advanced keyword search and the database on which the directory operates are both fully aware of international alphabets. While we are unable exhaustively to test and verify that searches using characters from non-English alphabets conform strictly to the conventions of the alphabet in question, there is no reason they should not work. If you are unsure whether the member you are looking for spells his or her name with an accent on a letter, in nearly all cases a search using the letter without an accent will produce the results you are looking for. That is to say if you are looking for someone named Josée entering the text Josee will suffice (as will Jos?e).

Examples

More examples of advanced keyword searches will follow soon.