This topic describes the properties listed in the Performance section of the Locator Properties dialog box.
Data cache size
This is the amount of RAM in megabytes the locator can allocate for caching the reference data of the locator. Geocoding will be faster as more data is cached in the memory when the data cache size is increased.
Follow the steps below to change the amount of memory allocated to caching reference data:
- On the Locator Properties dialog box, click the Performance tab on the left.
- Click the text box under the Data cache size option, and enter a new number in megabytes.
- Click OK to save the setting.
Presort input table by (fields)
This determines the sorting order, by field name, of the input table. If the locator is large, parts of the reference data in the locator are cached in the computer's memory for fast access. Therefore, it's more efficient to geocode a table that is sorted by geography, such as state name, ZIP Code, and city name, because the inputs can be compared faster with the data that is currently in the cache. On the other hand, if the records in the input table are not sorted and the inputs are scattered over a large area, it may take a longer time to geocode. The Presort input table by (fields) option allows the geocoding process to sort the records on the fly based on the specified fields.
- To change the presort field setting, click the Add button on the right of the Presort input table by (fields) table.
A fields context menu opens.
- Choose the fields you want to use for sorting in the Input fields list by checking the boxes associated with each field.
The field names in the list are in the order the records would be sorted. To change the sort order of the fields, click a field name in the list and use the Up arrow or Down arrow buttons to move the name up or down in the list.
Presort cache size (number of records)
This is the number of records in the presort cache. This option only applies if input table fields are specified in the Presort input table by (fields) option. A subset of data in the input table, defined by the presort cache size, is cached in the computer memory. The records are sorted and geocoded. The process repeats on the subsequent subsets of the data until the table is completely processed.
You can change the presort cache size on the Locator Properties dialog box by following the steps below.
- Click the text box under the Presort cache size (number of records) option, and enter the number of records that you want to cache for presorting.
- Click OK to save the setting.
Number of threads (file folder locators only)
This setting refers to the number of threads that are used for geocoding. To improve geocoding performance, it is recommended to use multiple threads. The capacity is determined by the number of processors available. You can set the property to Auto if you want it to be determined automatically.
Multithreading on a composite locator, which also needs to be in the same file folder as the participating locators, is determined by the setting of each participating locator. Each participating locator will create its own predefined number of threads.
Multithreading options are also honored for locators published as geocoding services running on ArcGIS for Server.
Maximum number of candidates
This is the maximum number of candidates to be returned in interactive match or rematch. Once the maximum number of candidates is returned, the search stops.
Maximum number of perfect candidates
This is the number of perfect match candidates (in other words, a score of 100) that are to be returned in interactive match or rematch. Once the maximum number of perfect candidates is returned, the search stops.
This value, in seconds, sets the time-out for the candidate search. For a locator that covers a very large area and contains many candidates of similar content, a common address without specific zone information, such as 100 Main Street, may take a longer time to search. Setting a search time-out limits the amount of time you spend waiting for a result, but since the time-out causes the search to stop prematurely, potential candidates may be overlooked, and the best location may not be found. The recommended practice is to provide more complete input information so the search can be more efficient.