Trouble with a composite _id

Hello!

I'm using the CouchDB _river to stream _changes from Big Couch to ES
0.20.2. I'm using full dynamic mapping for the _river.

The issue is that in CouchDB, we construct _ids in this fasion:
"66_DeliveryEventsSpec_messageid0_0_0_20130117T215701" Sometimes the _id
can have dashes (-) as well.

ES doesn't seem to like that _id setup. All the docs show correctly in the
index after being copied over by the _river and the _id field gets
populated from the CouchDB _id. But when I

http://localhost:9200/index_test/DeliveryEvent/66_DeliveryEventsSpec_messageid0_0_0_20130117T215701

I get

{
"_index": "index_test",
"_type": "DeliveryEvent",
"_id": "66_DeliveryEventsSpec_messageid0_0_0_20130117T215701",
"exists": false
}

Is that a limitation of ES or is there a setting where I can make the _id
field not analyzed? Any advice would be super. I'm hoping to perform script
updates on documents but without the _id working it's difficult.

Thanks!

--

Are you sure there's not something else going on? The _id field doesn't get
analyzed.

curl -XPUT
'http://localhost:9200/index_test/DeliveryEvent/66_DeliveryEventsSpec_messageid0_0_0_20130117T215701'
-d '{
"title":"test",
"body":"some random data"
}'

curl -XGET
'http://localhost:9200/index_test/DeliveryEvent/66_DeliveryEventsSpec_messageid0_0_0_20130117T215701?pretty=1'
{
"_index" : "index_test",
"_type" : "DeliveryEvent",
"_id" : "66_DeliveryEventsSpec_messageid0_0_0_20130117T215701",
"_version" : 1,
"exists" : true, "_source" : {
"title":"test",
"body":"some random data"
}
}

On Friday, January 18, 2013 11:56:03 AM UTC-5, JP Toto wrote:

Hello!

I'm using the CouchDB _river to stream _changes from Big Couch to ES
0.20.2. I'm using full dynamic mapping for the _river.

The issue is that in CouchDB, we construct _ids in this fasion:
"66_DeliveryEventsSpec_messageid0_0_0_20130117T215701" Sometimes the _id
can have dashes (-) as well.

ES doesn't seem to like that _id setup. All the docs show correctly in the
index after being copied over by the _river and the _id field gets
populated from the CouchDB _id. But when I

http://localhost:9200/index_test/DeliveryEvent/66_DeliveryEventsSpec_messageid0_0_0_20130117T215701

I get

{
"_index": "index_test",
"_type": "DeliveryEvent",
"_id": "66_DeliveryEventsSpec_messageid0_0_0_20130117T215701",
"exists": false
}

Is that a limitation of ES or is there a setting where I can make the _id
field not analyzed? Any advice would be super. I'm hoping to perform script
updates on documents but without the _id working it's difficult.

Thanks!

--

That must be it, then. It's gotta be something in my index. Thanks!

On Friday, January 18, 2013 3:12:19 PM UTC-5, egaumer wrote:

Are you sure there's not something else going on? The _id field doesn't
get analyzed.

curl -XPUT '
http://localhost:9200/index_test/DeliveryEvent/66_DeliveryEventsSpec_messageid0_0_0_20130117T215701'
-d '{
"title":"test",
"body":"some random data"
}'

curl -XGET '
http://localhost:9200/index_test/DeliveryEvent/66_DeliveryEventsSpec_messageid0_0_0_20130117T215701?pretty=1
'
{
"_index" : "index_test",
"_type" : "DeliveryEvent",
"_id" : "66_DeliveryEventsSpec_messageid0_0_0_20130117T215701",
"_version" : 1,
"exists" : true, "_source" : {
"title":"test",
"body":"some random data"
}
}

On Friday, January 18, 2013 11:56:03 AM UTC-5, JP Toto wrote:

Hello!

I'm using the CouchDB _river to stream _changes from Big Couch to ES
0.20.2. I'm using full dynamic mapping for the _river.

The issue is that in CouchDB, we construct _ids in this fasion:
"66_DeliveryEventsSpec_messageid0_0_0_20130117T215701" Sometimes the _id
can have dashes (-) as well.

ES doesn't seem to like that _id setup. All the docs show correctly in
the index after being copied over by the _river and the _id field gets
populated from the CouchDB _id. But when I

http://localhost:9200/index_test/DeliveryEvent/66_DeliveryEventsSpec_messageid0_0_0_20130117T215701

I get

{
"_index": "index_test",
"_type": "DeliveryEvent",
"_id": "66_DeliveryEventsSpec_messageid0_0_0_20130117T215701",
"exists": false
}

Is that a limitation of ES or is there a setting where I can make the _id
field not analyzed? Any advice would be super. I'm hoping to perform script
updates on documents but without the _id working it's difficult.

Thanks!

--