| alexander.behm | c576c60 | 2012-07-06 02:41:15 +0000 | [diff] [blame^] | 1 | drop dataverse test if exists; |
| 2 | create dataverse test; |
| 3 | use dataverse test; |
| 4 | |
| 5 | create type DBLPType as closed { |
| 6 | id: int32, |
| 7 | dblpid: string, |
| 8 | title: string, |
| 9 | authors: string, |
| 10 | misc: string |
| 11 | } |
| 12 | |
| 13 | create nodegroup group1 if not exists on nc1, nc2; |
| 14 | |
| 15 | create dataset DBLP(DBLPType) |
| 16 | partitioned by key id on group1; |
| 17 | |
| 18 | load dataset DBLP |
| 19 | using "edu.uci.ics.asterix.external.dataset.adapter.NCFileSystemAdapter" |
| 20 | (("path"="nc1://data/dblp-small/dblp-small-id.txt"),("format"="delimited-text"),("delimiter"=":")) pre-sorted; |
| 21 | |
| 22 | create index ngram_index on DBLP(title) type ngram(3); |
| 23 | |
| 24 | write output to nc1:"rttest/index-selection_inverted-index-ngram-jaccard.adm"; |
| 25 | |
| 26 | for $o in dataset('DBLP') |
| 27 | let $jacc := similarity-jaccard-check(gram-tokens($o.title, 3, false), gram-tokens("Transactions for Cooperative Environments", 3, false), 0.5f) |
| 28 | where $jacc[0] |
| 29 | return $o |