The merge operation is designed to combine two or more indexes into one index to accelerate the search and to simplify the work with indexes. When merging, only the index at which the Merge method was called is changed. This index as a result of the operation contains all the documents that were contained in all indexes together. The second index or index repository after the merge can be deleted to free up disk space. Since version 26.11, a merged index stays in its index repository, so its events reach the events of the repository after the merge as well; before, the merge moved the merged indexes out of their repository.
The MergeOptions class contains a property for specifying a cancellation object, as well as a property for specifying whether the operation will be performed asynchronously. By default, index merging operation is performed synchronously.
The Index class has method overloads for merging a single index and a whole index repository. Before performing a merge, checks are made to ensure that the indexes are of the same type. Merging is performed only on those indexes that have the same IndexType value as the index on which the Merge method is called.
Since version 26.11, a merge verifies the checksums of the index data that it reads. A merge that reads damaged data fails: the ErrorOccurred event reports that a file of the index is damaged, and the index keeps its state. Repair the damaged index with the CheckSegments method and repeat the merge (see Checking index integrity). Damaged posting lists of the whole current index are made again from its segments, and the ErrorOccurred event reports it as an error that is not critical.
Merging compatibility
Since version 26.10, only indexes that divide the text of documents into terms in the same way can be merged:
Before making any changes, the Merge method compares the current index with every merged index that contains documents: the UseStopWords, UseCharacterReplacements, ReplaceUnpairedSurrogates and SkipBlendedOnlyTerms settings, the alphabet, the tokenization dictionary, and also the stop words and the character replacements if the indexes use them. If a difference is found, the method throws InvalidOperationException, whose message names the differences, and no index is changed. This also happens in the asynchronous mode.
The alphabet, the tokenization dictionary, the stop word dictionary and the character replacement dictionary of the current index are not changed by merging. The other dictionaries of the merged indexes (for example, synonyms and homophones) are merged into the dictionaries of the current index. Since version 26.11, the synonyms and homophones of every word are kept as they are: after a merge, a word has the synonyms and homophones that the merged dictionaries give it. Before, the words of the groups that shared a word became synonyms of each other.
A merged index created by a previous version must be updated with the IndexUpdater class before merging (see Update index).
A merge is not performed if a merged index is performing an operation or has failed; the ErrorOccurred event reports it.
Since version 26.11, the default alphabet and tokenization dictionary of a new index have the letters of more scripts (see Character types), while an index keeps the alphabet that it was created with. Therefore, a new index is not merged with an index created by an earlier version or updated from one: the message of the exception names the Alphabet and the TokenizationDictionary. To merge them, give the new index the alphabets of the earlier index before documents are added to it: export the alphabet and the tokenization dictionary of the earlier index with the ExportDictionary method, and call the Clear method and then the ImportDictionary method for each of them in the new index. The example is presented on the Alphabet page; after that, the earlier index can be merged into the new one.
Merging example
The following example demonstrates the merging of two indexes.
C#
stringindexFolder1=@"c:\MyIndex1\";stringindexFolder2=@"c:\MyIndex2\";stringdocumentsFolder1=@"c:\MyDocuments1\";stringdocumentsFolder2=@"c:\MyDocuments2\";Indexindex1=newIndex(indexFolder1);// Creating index1index1.Add(documentsFolder1);// Indexing documentsIndexindex2=newIndex(indexFolder2);// Creating index2index2.Add(documentsFolder2);// Indexing documentsMergeOptionsoptions=newMergeOptions();options.Cancellation=newCancellation();// Creating cancellation object to be able to cancel the operationtry{// Merging index2 into index1// Files of index2 will not be changedindex1.Merge(index2,options);}catch(InvalidOperationExceptionexception){// The indexes divide text into terms differently; the message names the differencesConsole.WriteLine(exception.Message);}
More resources
GitHub examples
You may easily run the code from documentation articles and see the features in action in our GitHub examples: