(My) Problem with geo_bounding_box

Elastic Stack Elasticsearch
1

Hi,

I'm a new es user but love what I've seen so far. I'm using 0.90.0.

I have a simple mapping defined like this:

{
"mappings" : {
"user" : {
"properties" : {
"user_name" : {
"type" : "string",
"index_options" : "freqs",
"boost" : 3.0
},
"handle" : {
"type" : "string",
"index_options" : "docs"
},
"bio" : {
"type" : "string",
"index_options" : "freqs"
},
"location" : {
"type" : "geo_point",
"lat_lon" : true
}
}
},
"post" : {
"properties" : {
"text" : {
"type" : "string",
"index_options" : "freqs"
},
"location" : {
"type" : "geo_point",
"lat_lon" : true
}
}
}
}
}

I've loaded around 50k entries into the index.

I want to be able to perform searches based on text and optionally filtered
by a geo_bounding_box.

If I issue the following search, it works beautifully:

curl -XGET localhost:9200/wotzere/_search?pretty=1 -d '
{
"fields" : [ "_id", "_type", "location" ],
"from" : 0,
"size" : 20,
"query" : {
"match": {
"_all" : "chill ice lodge"
}
}
}'

It returns:

{
"took" : 5,
"timed_out" : false,
"_shards" : {
"total" : 5,
"successful" : 5,
"failed" : 0
},
"hits" : {
"total" : 1170,
"max_score" : 2.2864966,
"hits" : [ {
"_index" : "wotzere",
"_type" : "user",
"_id" : "80248",
"_score" : 2.2864966,
"fields" : {
"location" : [ "145.0450", "-37.8232" ]
}
}, {
"_index" : "wotzere",
"_type" : "user",
"_id" : "66694",
"_score" : 1.3186983,
"fields" : {
"location" : [ "144.9679", "-37.8205" ]
}
}, {
"_index" : "wotzere",
"_type" : "user",
"_id" : "72829",
"_score" : 0.76025414,
"fields" : {
"location" : [ "151.3003", "-32.7730" ]
}
}, {
"_index" : "wotzere",
"_type" : "user",
"_id" : "70022",
"_score" : 0.6828435,
"fields" : {
"location" : [ "150.2991", "-34.6497" ]
}
}, {
"_index" : "wotzere",
"_type" : "post",
"_id" : "138114",
"_score" : 0.52798617,
"fields" : {
"location" : [ "130.8392", "-12.4630" ]
}
}, {
:
} ]
}
}

For the life of me I can't get a geo_bounding_box filter to work - even on
its own. Would some kind soul please point me in the right direction by
incorporating a bounding box of (lat, long) tl => (-10, 110), br => (-44,
155) into my original search criteria?

One thing that's not clear to me. If I have a field called "location" in
both my types (user, post), do I need to specify both in separate
geo_bounding_box filters or can I do it with a single, unqualified filter
(sort of like what I'm doing with the "_all" of the original criteria)?

Thanks so much.

Steve

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

2

Hi Steve,

you can setup your geo_bounding_box filter by just adding a
geo_bounding_box filter:

curl -XGET 'http://127.0.0.1:9200/posts/_search?pretty=true' -d '{

    "fields" : [ "_id", "_type", "location" ],
    "from" : 0,
    "size" : 20,
    "query" : {
        "match": {
            "_all" : "chilling"
        }
    },

    "filter" : {
        "geo_bounding_box" : {
            "location" : {
                "top_left" : {
                    "lat" : -10.0,
                    "lon" : 110.0
                },
                "bottom_right" : {
                    "lat" : -44.0,
                    "lon" : 155
                }
            }
        }
    }

}'

I hope, this will help you,
Florian

On Monday, May 6, 2013 7:03:04 AM UTC+2, Steve Baldwin wrote:

Hi,

I'm a new es user but love what I've seen so far. I'm using 0.90.0.

I have a simple mapping defined like this:

{
"mappings" : {
"user" : {
"properties" : {
"user_name" : {
"type" : "string",
"index_options" : "freqs",
"boost" : 3.0
},
"handle" : {
"type" : "string",
"index_options" : "docs"
},
"bio" : {
"type" : "string",
"index_options" : "freqs"
},
"location" : {
"type" : "geo_point",
"lat_lon" : true
}
}
},
"post" : {
"properties" : {
"text" : {
"type" : "string",
"index_options" : "freqs"
},
"location" : {
"type" : "geo_point",
"lat_lon" : true
}
}
}
}
}

I've loaded around 50k entries into the index.

I want to be able to perform searches based on text and optionally
filtered by a geo_bounding_box.

If I issue the following search, it works beautifully:

curl -XGET localhost:9200/wotzere/_search?pretty=1 -d '
{
"fields" : [ "_id", "_type", "location" ],
"from" : 0,
"size" : 20,
"query" : {
"match": {
"_all" : "chill ice lodge"
}
}
}'

It returns:

{
"took" : 5,
"timed_out" : false,
"_shards" : {
"total" : 5,
"successful" : 5,
"failed" : 0
},
"hits" : {
"total" : 1170,
"max_score" : 2.2864966,
"hits" : [ {
"_index" : "wotzere",
"_type" : "user",
"_id" : "80248",
"_score" : 2.2864966,
"fields" : {
"location" : [ "145.0450", "-37.8232" ]
}
}, {
"_index" : "wotzere",
"_type" : "user",
"_id" : "66694",
"_score" : 1.3186983,
"fields" : {
"location" : [ "144.9679", "-37.8205" ]
}
}, {
"_index" : "wotzere",
"_type" : "user",
"_id" : "72829",
"_score" : 0.76025414,
"fields" : {
"location" : [ "151.3003", "-32.7730" ]
}
}, {
"_index" : "wotzere",
"_type" : "user",
"_id" : "70022",
"_score" : 0.6828435,
"fields" : {
"location" : [ "150.2991", "-34.6497" ]
}
}, {
"_index" : "wotzere",
"_type" : "post",
"_id" : "138114",
"_score" : 0.52798617,
"fields" : {
"location" : [ "130.8392", "-12.4630" ]
}
}, {
:
} ]
}
}

For the life of me I can't get a geo_bounding_box filter to work - even on
its own. Would some kind soul please point me in the right direction by
incorporating a bounding box of (lat, long) tl => (-10, 110), br => (-44,
155) into my original search criteria?

One thing that's not clear to me. If I have a field called "location" in
both my types (user, post), do I need to specify both in separate
geo_bounding_box filters or can I do it with a single, unqualified filter
(sort of like what I'm doing with the "_all" of the original criteria)?

Thanks so much.

Steve

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

3

Fantastic !! That was very helpful.

Thank you very much Florian.

Steve

On Monday, May 6, 2013 9:42:05 PM UTC+10, Florian Schilling wrote:

Hi Steve,

you can setup your geo_bounding_box filter by just adding a
geo_bounding_box filter:

curl -XGET 'http://127.0.0.1:9200/posts/_search?pretty=true' -d '{

    "fields" : [ "_id", "_type", "location" ],
    "from" : 0,
    "size" : 20,
    "query" : {
        "match": {
            "_all" : "chilling"
        }
    },

    "filter" : {
        "geo_bounding_box" : {
            "location" : {
                "top_left" : {
                    "lat" : -10.0,
                    "lon" : 110.0
                },
                "bottom_right" : {
                    "lat" : -44.0,
                    "lon" : 155
                }
            }
        }
    }

}'

I hope, this will help you,
Florian

On Monday, May 6, 2013 7:03:04 AM UTC+2, Steve Baldwin wrote:

Hi,

I'm a new es user but love what I've seen so far. I'm using 0.90.0.

I have a simple mapping defined like this:

{
"mappings" : {
"user" : {
"properties" : {
"user_name" : {
"type" : "string",
"index_options" : "freqs",
"boost" : 3.0
},
"handle" : {
"type" : "string",
"index_options" : "docs"
},
"bio" : {
"type" : "string",
"index_options" : "freqs"
},
"location" : {
"type" : "geo_point",
"lat_lon" : true
}
}
},
"post" : {
"properties" : {
"text" : {
"type" : "string",
"index_options" : "freqs"
},
"location" : {
"type" : "geo_point",
"lat_lon" : true
}
}
}
}
}

I've loaded around 50k entries into the index.

I want to be able to perform searches based on text and optionally
filtered by a geo_bounding_box.

If I issue the following search, it works beautifully:

curl -XGET localhost:9200/wotzere/_search?pretty=1 -d '
{
"fields" : [ "_id", "_type", "location" ],
"from" : 0,
"size" : 20,
"query" : {
"match": {
"_all" : "chill ice lodge"
}
}
}'

It returns:

{
"took" : 5,
"timed_out" : false,
"_shards" : {
"total" : 5,
"successful" : 5,
"failed" : 0
},
"hits" : {
"total" : 1170,
"max_score" : 2.2864966,
"hits" : [ {
"_index" : "wotzere",
"_type" : "user",
"_id" : "80248",
"_score" : 2.2864966,
"fields" : {
"location" : [ "145.0450", "-37.8232" ]
}
}, {
"_index" : "wotzere",
"_type" : "user",
"_id" : "66694",
"_score" : 1.3186983,
"fields" : {
"location" : [ "144.9679", "-37.8205" ]
}
}, {
"_index" : "wotzere",
"_type" : "user",
"_id" : "72829",
"_score" : 0.76025414,
"fields" : {
"location" : [ "151.3003", "-32.7730" ]
}
}, {
"_index" : "wotzere",
"_type" : "user",
"_id" : "70022",
"_score" : 0.6828435,
"fields" : {
"location" : [ "150.2991", "-34.6497" ]
}
}, {
"_index" : "wotzere",
"_type" : "post",
"_id" : "138114",
"_score" : 0.52798617,
"fields" : {
"location" : [ "130.8392", "-12.4630" ]
}
}, {
:
} ]
}
}

For the life of me I can't get a geo_bounding_box filter to work - even
on its own. Would some kind soul please point me in the right direction by
incorporating a bounding box of (lat, long) tl => (-10, 110), br => (-44,
155) into my original search criteria?

One thing that's not clear to me. If I have a field called "location" in
both my types (user, post), do I need to specify both in separate
geo_bounding_box filters or can I do it with a single, unqualified filter
(sort of like what I'm doing with the "_all" of the original criteria)?

Thanks so much.

Steve

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.