Highlighting Creates Incorrect JSON Format for Response


Using Elasticsearch 1.6.0 and am getting some weird behavior from highlighting. When I add highlighting to my queries, all responses have an extra comma (",") between the "_source" and the "highlight" return fields, except for those that are pretty printed to the command line (pretty printing to a file also causes the double comma). The data I am currently searching over is confidential so I cannot show an exact example of the output, but it is in essence the following:

{"_index":"index","_type":"type","_id":"id","_score":"2","_source":{...source text...LASTFIELD:LASTVALUE},
,"highlight" : {"FIELDHIGHLIGHTED" :"..."}},next_document

The "highlight" field is always on the next line from the document it is related to, with a comma at the end of the document's line and at the beginning of the highlight's line.

Any ideas what might be going on? Thanks!


Issue resolved, it was from some improperly formatted documents I had indexed.

Never, ever, ever try to index anything that ends in a comma. They will index and search fine, but the "_source" field will contain a trailing comma which can mess up a ton of JSON parsing.