This is related to my earlier questions
- Spatial query with sub-select (I figured ths one out)
- OrientDB spatial query to find all pairs within X km of each other (still looking for a useful answer)
In response to (2), I am looking at modifying my nazca geoglyph dataset to use the WKT version to be consistent with the newer OrientDB 2.2.x Spatial Index functionality.
My input CSV file, nazca_lines_wkt.csv
is this:
Name,Location
Hummingbird,POINT(-75.148892 -14.692131)
Monkey,POINT(-75.138532 -14.706940)
Condor,POINT(-75.126208 -14.697444)
Spider,POINT(-75.122381 -14.694145)
Spiral,POINT(-75.122746 -14.688277)
Hands,POINT(-75.113881 -14.694459)
Tree,POINT(-75.114520 -14.693898)
Astronaut,POINT(-75.079755 -14.745222)
Dog,POINT(-75.130788 -14.706401)
Wing,POINT(-75.100385 -14.680309)
Parrot,POINT(-75.107498 -14.689463)
I create an empty PLOCAL database, nazca-wkt.orientdb
and define a GeoGlyphWKT vertex class:
CREATE DATABASE PLOCAL:nazca-wkt.orientdb admin admin plocal graph
CREATE CLASS GeoGlyphWKT EXTENDS V
CREATE PROPERTY GeoGlyphWKT.Name STRING
CREATE PROPERTY GeoGlyphWKT.Location EMBEDDED OPoint
CREATE PROPERTY GeoGlyphWKT.Tag EMBEDDEDSET STRING
I have two .json files that I use for the oetl script:
nazca_lines_wkt.json
{
"config": {
"log": "info",
"fileDirectory": "./",
"fileName": "nazca_lines_wkt.csv"
}
}
commonGeoGlyphWKT.json
{
"begin": [ { "let": { "name": "$filePath", "expression": "$fileDirectory.append($fileName )" } } ],
"config": { "log": "debug" },
"source": { "file": { "path": "$filePath" } },
"extractor":
{
"csv": { "ignoreEmptyLines": true,
"nullValue": "N/A",
"separator": ",",
"columnsOnFirstLine": true,
"dateFormat": "yyyy-MM-dd"
}
},
"transformers": [ { "vertex": { "class": "GeoGlyphWKT" } } ],
"loader": {
"orientdb": {
"dbURL": "plocal:nazca-wkt.orientdb",
"dbType": "graph",
"batchCommit": 1000
}
}
}
I run oetl using this command:
$ oetl.sh commonGeoGlyphWKT.json nazca_lines_wkt.json
but this fails with the following output:
$ oetl.sh commonGeoGlyphWKT.json nazca_lines_wkt.json
OrientDB etl v.2.2.13 (build 2.2.x@r90d7caa1e4af3fad86594e592c64dc1202558ab1; 2016-11-15 12:04:05+0000) www.orientdb.com
BEGIN ETL PROCESSOR
[file] INFO Reading from file ./nazca_lines_wkt.csv with encoding UTF-8
Started execution with 1 worker threads
Error in Pipeline execution: com.orientechnologies.orient.core.exception.OValidationException: impossible to convert value of field "Location"
DB name="nazca-wkt.orientdb"
ETL process has problem: java.util.concurrent.ExecutionException: com.orientechnologies.orient.core.exception.OValidationException: impossible to convert value of field "Location"
DB name="nazca-wkt.orientdb"
END ETL PROCESSOR
+ extracted 9 rows (0 rows/sec) - 9 rows -> loaded 0 vertices (0 vertices/sec) Total time: 16ms [0 warnings, 1 errors]
I'm sure it's something silly that I'm missing... has anyone been able to import CSV files containing WKT strings for points, polygons, etc using ETL?
Any help is appreciated!