For the same data with data format integer and long works in 5.6 ES but not in 7.6 ES

Hi,

Thank you in advance,

I am reindexing data from 5.6 ES to 7.6 ES cluster. Before reindexing, we have created an index same name and mapping as in 5.6 es cluster. But When we started reindexing, the data values are too big for integer and long. Although, the same data in 5.6 is not throwing any error.

Note: We are using spark to ingest data into 5.6 ES cluster.

Is it possible, that you configured in the mapping, that in case of an error, the document should be indexed anyway?

See the ignore_malformed mapping parameter.

Hi @spinscale

Thank you for the reply.

No i there is no ignore_malformed field in the mapping.

image

can you share exceptions and sample documetns to reproduce the issue?

Thanks!