Advanced search

From Topicscape
Jump to navigation Jump to search
Previous: Search

(Go to:Topicscape Pro User Guide - contents list)

Advanced Searches offer many options:

  • Topic and occurrence information searches
    This is likely to be the most common type of search you do.
    What it does for you: Helps you find items based on their name, and information that you have added, including flags and colors.
    Searches for: Keywords or a phrase.
    Where it looks: This type of search can look in topic names, occurrence names, filenames, as well as description, author, source and authority information that you have added (metadata). There are options to include or exclude these categories of information, and to select by date and file type.
    There is also an option to just search down one branch, starting at its top-most topic.
    How to start it: Just type your search keywords (or phrase in double quotes) and select where you want to look. If you are editing text in the details panel, naturally, you cannot search until you have finished.
    There are two search panels for this type of search - a comprehensive one that gives access to all the selection features, and a simple one that just shows the controls that you will need for most quick searches. The type of panel you see can be chosen from a button on the panel itself.
    For more detail, see: Initiating an advanced search.
  • File-content search
    What it does for you: Extends your searching capability to file content.
    What it does: Examines files themselves;
    Searches for: A phrase that you specify.
    Where it looks: inside the files attached to occurrences in the currently-open Topicscape.
    How to start it: Just type your search phrase (with or without double quotes -- it will treat this as a phrase search either way) and select the file content option. You can also limit the search to one type of file (file extension) and specific date ranges if you wish.
    For more detail, see: File Content searches.
  • Association-type search
    *What it does for you: Allows you to limit the search to any phrases you may have added to describe how topics are associated;
    Searches for: Keywords or phrase;
    Where it looks: In just the association type;
    How to start it: Just type your search keywords (or phrase in double quotes) and tick the association type checkbox.
    For more detail, see: Association types - capturing knowledge.
  • Topic-limited search
    What it does for you: Allows you to limit the results of a search, by specifying the part topic in which it may appear. Provides the above two types of Advanced Search functions, but confined to a limited part of a Topicscape that you select;
    Searches for: Keywords, or a phrase, or file content, according to options that you select;
    Where it looks: Within a single branch of a Topicscape, starting at a topic that you specify and working down through the children, grandchildren, etc.. It examines either metadata as described above (under 'Topic and occurrence information searches') or file content, depending on options that you select;
    How to start it: right-click a topic and select "Search this topic & descendants".
    The search panel has an orange frame around the text box to confirm that the search will be limited to a specified group of topics.
    For more detail, see: Group search.


Initiating an advanced search

ASP.jpg

To start on advanced search you just start typing when you are in the Topic Center or looking at the Map or 'Scape. The search panel pops up and whatever you've already keyed in will appear in the search box. It will be forced to lower case. Most searches in Topicscape ignore case so if a match is found it is independent of capital letters, so "this", "THIS" and "This" are treated as equal. The exception is in a prefix search (when a wildcard, *, is used - see below), when all lower case is required.

The advanced search generally uses keywords: type in the words you want to find and press Enter for the easiest form of search. That will look for all of the words you typed (technically, the search words will be 'ANDed') and the order in which they are found will be ignored. For completeness, note that it will not usually search for words carrying little meaning, like "the", "and", "a", "but". In Topicscape, these are referred to as noisewords and a standard set is installed automatically. You can change them (see Noisewords)

Often, you will just want a simple search in Topic names and/or occurrence names. In that case, you may find you prefer working with the Simple Search Panel, which hides the functions you are not using. Switching backward and forward between these panels is done by a single click.

Where it looks for your search target will depend on the options you selected in the Advanced Search Panel. It will look by default in all topic names and all occurrence and file names. You can change the defaults to your own preference as described in Set your own default search options, the last item in the table below). There are other options to include Descriptions of topics and occurrences, and origins of occurrences in the search. Origins are the source, author's name and authority, if any, of an occurrence.

As you fill in the keywords or phrase and select options, the blue text in the white Search statement box will change to provide a description of the search you are building.

The search finds exact keyword matches only. To search for prefixes of words (including plurals) you can use an asterisk like this: "ref*". This will return hits for all words starting with "ref". See Wildcard option in the Glossary for a description of the types of wildcards supported.

You can narrow the search, if you wish: By selecting a file extension, you can look for occurrences that include only a specific type of file; by selecting a type of date, you can limit your search to, for example, topics you've visited in the last 3 days, or MS Excel files you've opened (via Topicscape) in the last week.

You can further narrow the search by limiting it to the group under a specific topic: Its strict descendants -- its children, grandchildren and so on. Please see Group search.

Topicscape looks at the check boxes to decide where it should look for the target words or phrase. All of the keywords must appear in at least one of the locations checked for a hit to be found. In other words if keyword one is in the topic name but keyword two is not, and keyword two is in the occurrence name but keyword one is not, no hit will be recorded. The check boxes are:

topic name                    Search in the topic name
all topic details As well as the topic name, this includes the topic description in the search.
occurrence (and file name)
all occurrence and file details As well as the occurrence name and filename, this includes the occurrence description, author, source and authority in the search.
file content This looks in the files that are linked to occurrences. It searches the file without knowing the internal structure expected by the application that made it. With many applications, this means that the search will be useful but files made by some software, for example Acrobat (pdf) files and PowerPoint (ppt), will cannot be searched usefully.
use Google Desktop If you have Google Desktop installed, you can make Topicscape use that instead of the slower, direct contents search, to search file content by checking this checkbox.
association type you can make the search look in association types with this checkbox.
find noisewords too If there is a noiseword ('the', 'an', 'it', for example) in the search string, checking this option will require the noisewords to be found in one of the target areas.
Set your own default search options There is a button marked 'Set' that determines how the checkboxes will be set each time you open the Advanced Search panel. Press that button and the settings of the six checkboxes above it will be remembered every time you open the panel.

Re-execute previous searches

Re-executeSearch.jpg

The text box where you type the search term in the Advanced Search Panel remembers the history of searching. Click on the right hand down arrow and you will see search terms that you have used before that start with the characters you have typed so far.

There is a new button in the search panel to clear search history. To delete a single item from the search history, use the down cursor key to highlight it (but don't press Enter) and then press Delete.

Searching in topic groups

You can do an advanced search focusing on just a specific topic and its children, grand-children, etc. by right-clicking the topic and selecting "Search this topic & its descendants". The text search box is orange to remind you that this is not a full Topicscape search.

Filtering search results

You can select the items that are to appear in the Hit List by date range and file type (extension) using the two tabs at the foot of the Advanced Search panel.

Once the search results are shown you can apply other filters using the Lists control bar, as you examine the results.

You can limit a search to just flagged topics, just un-flagged topics or include all topics in the search.

You can do a similar thing for occurrences as mentioned above - select whether the search should be limited to specific colors, or select whether to search only occurrences not marked "Finished with", or just the "Finished with" ones.

You can combine these selections with other criteria, like date or file type to narrow a search even more.

File Content searches - Topicscape's own searcher

Topicscape's file content search will look through the Topicscape's own folder, and in files linked through a 'link' occurrence. If you are working with a Topicscape called ThisOne, it might look in C:\Documents and Settings\Jan\My Topicscapes\ThisOne\. You can choose date ranges and filetypes to limit the search.

The search looks for plain text, without any processing for specific types of files. It supports search on Word and Excel files but will likely not give 100% accuracy if MS Word's 'Track changes' feature is in use, and the phrase you are searching for had been altered. When OLE objects are saved with paste special, linked and an occurrence has been made for you, Topicscape will not find content that is inside the paste special boundary.

When the file-content search is searching HTML and emails it ignores tags (e.g. <BODY>) but not invisible data enclosed by tags. For example, it would find the xxxx in <Script> xxxx </SCRIPT> if xxxx were part of the search string.

For now, Topicscape cannot support reliable searching on Acrobat (pdf), PowerPoint (ppt) or archive (zip, etc.) files. If you need this, then you should consider one of the (free) alternatives described below.

Google Desktop can speed "file-content" searches

If you have Google Desktop installed, Topicscape gives you the option of using this for faster file-content searching. The first time you ask to do so, GDS will ask if you want to install Topicscape as a plug in. You must allow it to do so if you want this option for your searches. Google Desktop does not take part in other forms of Topicscape searching.

An item in the Options panel allows you to make Google Desktop the default file content searcher, or not. If Google Desktop is not installed, that item will be grayed-out. When you do a file-content search, you can accept the default that you chose in the Options panel, or you can override it.

Limitations to remember

1. Google Desktop will only provide results if your My Topicscapes folder is included in its indexes.

2. If you introduce many files, or Google Desktop still has many files to index (for example, if you have just installed it), you may not get the results you expect. You will have to wait for the indexing process to include new files introduced to Topicscape before you will be able to search their contents successfully.

3. When Google Desktop has already indexed a file outside Topicscape and you move it into Topicscape, you may experience problems. This relates to what appears to be a bug that we have reported on the Google Desktop forum, but to which we have no reply yet. Google Desktop itself can find the file. It provides a correct link, however its preview and cache entries are broken - they still point at the first location. More importantly, when Topicscape asks Google Desktop to provide the information through its Application Program Interface (API) it returns the wrong address (the first location), so Topicscape determines on the basis of this incorrect information that it is not in Topicscape's control and does not report its presence. Until this is resolved, the only solution is to rebuild your Google Desktop indexes from time to time (or switch to Windows Desktop Search). Topicscape can only query Google Desktop, it cannot change its indexes.

Windows Desktop Search can speed "file-content" searches

If you have Windows Desktop Search installed, Topicscape gives you the option of using this for faster file-content searching. Windows Desktop Search does not take part in other forms of Topicscape searching. An item in the Options panel allows you to switch the use of Windows Desktop Search on or off. If Windows Desktop Search is not installed, that item will be grayed-out.

An item in the Options panel allows you to make Windows Desktop Search the default file content searcher, or not. If Windows Desktop Search is not installed, that item will be grayed-out. When you do a file-content search, you can accept the default that you chose in the Options panel, or you can override it.

Windows Desktop Search and Google Desktop both installed

If both are installed, both will be accessible. Windows Desktop Search searches in .mht files, whereas Google Desktop does not.

File content word / phrase / string handling

The different file content searches handle the words searched for in different ways as described below.

Topicscape

This will report phrase hits, not keyword matches. This means that if you search for "meaningful relationships" and a document contains only "relationships meaningful", no hit will be reported. In a Topicscape file content search, the search result will be the same whether you search for a phrase in double quotes or in no quotes.

Technically, the search is a string search. That means that even if you search for "ningful relationshi", and a file contains "meaningful relationships", a match will be reported.

Google Desktop (GD)

This will report keyword matches. This means that if you search for "meaningful relationships" and a document contains only "relationships meaningful", a match will still be reported. In a GD file content search, the search result will be the same whether you search for a phrase in double quotes or in no quotes.

Windows Desktop Search (WDS)

This will report keyword matches and prefix matches. This means that if you search for "meaningful relationships" and a document contains only "relationships meaningful", a match will be reported. Unlike GD, if a searched document contains "meaningful relationships" and you search with WDS for "mean relations" a hit will still be reported. Searching for "ningful ationshi" will produce no results in such a case (unless the 'words' "ningful relationshi" themselves are actually present).

Phrase Searches

This relates to the regular Topicscape searches, searches in the topic, occurrence and file names, and description, author, etc., not the file content searches just discussed. Search for complete phrases by enclosing them in quotation marks. Words enclosed in double quotes ("poison ivy") will appear together in all results exactly as you have entered them. You may not use phrase searches with keyword searches. So to look for "poison ivy" cure is not supported. A phrase search is recognized by the opening double quote.

Any pattern matching symbols (wildcards) such as '*' or '?' appearing inside quotes will be taken literally, not treated as wildcard searches. For example, if the search phrase is "poison* ivy", then a topic named "poisonous ivy" will not result in a match. Nor will "poison ivy" -- the actual character * will have to be found in a searched item before it will be returned as a hit.

Phrases must begin and end in a double quote. The first quote must be the first character in the 'Search on' box. The final quote must not be omitted.

A phrase search of one word (for example "graph") will be treated as a regular keyword search for the word 'graph'. It will not match 'graphics'. In contrast, a phrase search for "richly-colored graph" would register a hit on a topic or occurrence name containing 'richly-colored graphics'.

If you enter a phrase in quotes and then select 'file contents', the search will be done as if the quotes were not present. In other words hits will be registered on documents containing the phrase whether it has double quotes around it or not.


Next: Local search

For an exciting free guide to all things visual, visit the
Visual Thinking Center
     'G+'