You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Imagine you have a collection of large JSON or XML documents, and you want to run queries over them that just grab a small subset of the data. You don't want to fully parse each document for each query, so you should index the documents somehow, but they all have different tree structures, and it's not clear what an index would look like. This paper shows how to augment each document with a small amount of data that makes it very fast to search inside them.
The text was updated successfully, but these errors were encountered:
http://www.di.unipi.it/~ottavian/files/semi_index_cikm.pdf
Imagine you have a collection of large JSON or XML documents, and you want to run queries over them that just grab a small subset of the data. You don't want to fully parse each document for each query, so you should index the documents somehow, but they all have different tree structures, and it's not clear what an index would look like. This paper shows how to augment each document with a small amount of data that makes it very fast to search inside them.
The text was updated successfully, but these errors were encountered: