Preparing data for replication

This topic applies to ArcGIS for Desktop Standard and ArcGIS for Desktop Advanced only.

The following steps should be considered when preparing data for replication:

Determine the datasets to replicate

Geodatabase replication allows you to replicate all datasets or a subset of datasets in your geodatabase. To be replicated, these datasets must meet the following requirements:

• The database user must have write access to the data.
• All data must be registered as fully versioned; it cannot be versioned with the option to move edits to base.

With checkout replicas and one-way child-to-parent replicas, you have the option of checking out versioned or nonversioned data.

Two-way and one-way parent-to-child replicas have these additional requirements:

• Each dataset must have a GlobalID column. This column is used to maintain row uniqueness across geodatabases.
• All spatial data must be stored in a high-precision spatial reference.

If one-way replication using archiving is going to be used, and the data is not already enabled for archiving, it will be enabled during the replica creation process.

Any dataset not meeting these requirements will not be included in the replica.

The list of data to replicate automatically expands to include dependent datasets. For example, all feature classes in a geometric network, topology, or feature dataset are included if just one feature class in the network, topology, or feature dataset is selected for replication.

Define the data to be replicated

For each dataset, you can choose to replicate all data, a subset of rows, or even just the schema. Plan on replicating an appropriate amount of data for your needs. Consider the lifetime of the replica and make sure your requirements are covered.

Replica creation determines the data to replicate using two mechanisms—filters and relationship classes.

There are three types of filters:

• Spatial: Geometry is used to determine the area to replicate.
• Selections: Data is replicated based on selection sets of individual feature classes and tables.
• Querydefs: Definition queries are applied to individual feature classes and tables.

When replicating in ArcMap, the spatial filter is determined by the current view extent of the ArcMap document or the boundary of a currently selected graphic. Features intersecting this filter are included. Definition queries and selections on individual layers and tables are also applied. If more than one filter is used, the intersection of all filters is applied.

Once data has been added to the replica from processing the filters, the relationship class logic is applied. Here, for each dataset involved in a relationship class, additional rows are added if they are related to the data already in the replica. See Replication and related data for more information.

The following is a list of the types of data for which additional rules and behaviors are applied when creating replicas. Review the topics that are appropriate for your data:

Geometric networksTopologyRaster dataTerrainsNetwork datasetsRepresentations

Metadata for the data you choose to replicate is copied during the replica creation process. However, changes to the metadata are not applied during replica synchronization.

Naming replicated datasets

When replicating data through the Create Replica wizard, you can choose to change the name of replicated datasets. This is done in the Advanced Create Replica Options dialog box of the wizard.

If you are replicating data or schema only, then your naming options for datasets are to leave the existing name or type in a new name for the dataset. For example, if you had data on the parent named US_Counties but were only interested in replicating the counties in California, you may want to change the name to California_Counties on the child instead of leaving the original name.

If you are replicating data with the option to register existing data only, you can either leave the original name of the dataset or choose from a drop-down list of available datasets. When you choose a dataset from the list, you are effectively mapping the dataset in the parent to the chosen dataset on the child (or vice versa if you are creating a one-way child-to-parent replica). For example, if you had two geodatabases, each containing similar datasets for U.S. counties, but these datasets had different naming conventions in their respective geodatabases, you could map datasets in the parent replica to the appropriate datasets in the child replica. For example, CA_Counties in the parent geodatabase could map to California_Counties in the child geodatabase.

An example of replicating data

The following electric utilities' maintenance work orders will help illustrate some of the default behavior for replicating data.

A maintenance crew is preparing to inspect some of the electric utilities in a residential area. To do some field editing, the crew needs to replicate that part of the electric network that covers this residential area. To start the replication process, the spatial extent of the inspection area is identified using a spatial filter (in this case, the extent is determined by a selected graphic).

The crew is to concentrate on cables that have been insulated with a particular material. To identify these cables, a query is applied to the relevant dataset.

Finally, as each maintenance crew can expect to visit only a certain number of properties in a day, the homes in one residential block are identified by a selection based on property numbers.

The selected features, features identified by a definition query, and features that intersect the chosen spatial extent will be replicated. Some additional network features have also been included. How geometric networks are replicated is explained in greater detail in Replication and geometric networks.