I get the following exception when i try to index a document.
Caused by: org.elasticsearch.index.mapper.MapperParsingException: failed to parse date field [J], tried both date format [dateOptionalTime], and timestamp number with locale []
and that is all correct, but is it possible to ignore the exception and index the rest of the fields?
I think this can be done without re-indexing by adding the "ignore_malformed" flag to the mapping for the field. Here's what I tried and it worked:
POST events
{
"mappings" : {
"dates" : {
"properties" : {
"lenient_date" : {
"type" : "date"
}
}
}
}
}
PUT events/dates/1
{
"lenient_date" : "1/32/2014"
}
Result: parsing error (as expected)
{
"error": "MapperParsingException[failed to parse [lenient_date]]; nested: MapperParsingException[failed to parse date field [1/32/2014], tried both date format [dateOptionalTime], and timestamp number with locale []]; nested: IllegalArgumentException[Invalid format: \"1/32/2014\" is malformed at \"/32/2014\"]; ",
"status": 400
}
GET events/dates/1 Result: date not found (as expected)
{
"_index": "events",
"_type": "dates",
"_id": "1",
"found": false
}
PUT events/dates/_mapping
{
"dates" : {
"properties" : {
"lenient_date" : {
"type" : "date",
"ignore_malformed" : true
}
}
}
}
PUT events/dates/1
{
"lenient_date" : "1/32/2014"
}
GET events/dates/1
Result: now works
{
"_index": "events",
"_type": "dates",
"_id": "1",
"_version": 1,
"found": true,
"_source": {
"lenient_date": "1/32/2014"
}
}
I'm running ES version 1.3.2. I have not checked how the updated mapping affects sorting and filtering.
Actually, you get this exception either because you mapped this field as a date or ElasticSearch dynamically identified this field as a date (the first document you used to create the index had a date value for this field).
You can't ignore this exception and index the other information as the document doesn't match the defined mapping.
There are several cases :
In fact, you don't have date values in this field : you can fix the mapping to what you really need.
You have date values in this field, but it's not a valid date according to the elasticsearch date formats : you can add this format to the one used by default (using dynamic_date_formats) or define the specific format of this field.
You have sometimes date values, sometimes string values : if it suits your need, you could set the mapping to string to prevent this exception from occuring. You won't get parsing exceptions as it won't get parsed as a date but it won't be possible to perform a range query on this field for example.
