Read index

Purpose: Search/update an index.

 read-index <index> \
     ( equal <search key> | lesser <search key> | greater <search key> | \
         lesser-equal <search key> | greater-equal <search key> | \
         min-key | max-key ) \
     [ value <value> ] \
     [ update-value <update value> ] \
     [ key <key> ] \
     [ status <status> ] \
     [ new-cursor <cursor> ]

Copied!

read-index will search <index> (created with new-index) for a node with the string key that is:

equal to <search key> ("equal" clause)
lesser than <search key> ("lesser" clause)
greater than <search key> ("greater" clause)
lesser or equal than <search key> ("lesser-equal" clause)
greater or equal than <search key> ("greater-equal" clause)
a minimum key in the index ("min-key" clause)
a maximum key in the index ("max-key" clause)

The <status> in "status" clause will be GG_OKAY if a key conforming to one of these criteria is found, and GG_ERR_EXIST if not.

If a key is found, the value associated with the key can be obtained with "value" clause in <value>; an existing key used to originally insert this value into the index can be obtained with "key" clause in string <key>. If a key is not found, both <value> and <key> are unchanged.

You can update the value associated with a found key with "update-value" clause by specifying <update value> string. This update is performed after <value> has been retrieved, allowing you to obtain the previous value in the same statement.

If you'd like to iterate the ordered list of keys in an index, create a <cursor> by using "new-cursor" clause, in which case <cursor> will be positioned on a found index node. See use-cursor for more on using cursors. Cursors are useful in range searches; typically you'd find a key that is an upper or lower bound of a range and then keep iterating to a lesser or greater value until some criteria is met, such as when the opposite bound is found. Gliimly indexes are by default constructed so that such iterations are O(1) in complexity, meaning each is a single index node access (see new-index).

Examples

In this example, a million key/value pairs are inserted into an index, and then each of them is searched for and then displayed back (see write-index for more on inserting into a index). Both the key and the data are a numerical value of a key:

 %% /index-example public

     new-index myindex key-as "positive integer" // create new index

     set-number i
     start-loop use i start-with 0 repeat 1000000
         number-string i to key
         set-string data=key
         write-index myindex key (key) value data // insert key/data to the index
     end-loop

     start-loop use i start-with 0 repeat 1000000
         number-string i to key
         // search index for each key previously inserted
         read-index myindex equal (key) status st value data
         if-true st not-equal GG_OKAY
             @Could not find key <<p-out key>>
         else-if
             @Found data <<p-out data>> associated with key <<p-out key>>
         end-if
         delete-string key
     end-loop
 %%

Copied!