When retrieving data from the Ardor api, why is the data returned not in chronological order?
asked Oct 10, 2018 at 20:08
-
1Which API call are you doing exactly? It's supposed to be in reverse chronological order: nxtwiki.org/wiki/The_Nxt_API#Get_All_Tagged_Data– MrVCommented Oct 10, 2018 at 20:35
-
searchTaggedData– Patoshi パトシCommented Oct 10, 2018 at 20:36
-
It's not in any order it seems: prnt.sc/l4ln1k– Patoshi パトシCommented Oct 10, 2018 at 20:42
Add a comment
|
2 Answers
The results are first ordered by the score which Apache Lucene assigns to them . If two results get same score, they are ordered by blockTimestamp, descending (what happens with the 2nd, 3rd and 4th result in your data).
Apache Lucene is configured with the default BM25Similarity with parameters k1 = 1.2 b = 0.75, i.e. Ardor uses the default scoring function. More details about that function can be found here. The query parser is configured with default operator AND, so if your query contains two words, they must both be present in order for the data to be included in the result
From petko:
They are ordered by ScoreDoc.score - see https://lucene.apache.org/core/7_2_1/core/index.html?org/apache/lucene/search/ScoreDoc.html
