op.xpath works slowly for multiple paths

Is there any other option to extract columns from result than using the op.xpath?

For my code:

var data = op
  .fromSearchDocs(cts.collectionQuery('xxx'))
  .limit(100)
  .select(columns.map(x => op.as(x, op.xpath('doc', '/' + x))))
  .result()
  .toArray()

  data

This select:

.select(columns.map(x => op.as(x, op.xpath('doc', '/' + x))))

Increases the execution time of my code 10 times. The number of columns I want to extract is up to 500 and it is dynamic.

In other hand extracting data with jsearch:

import jsearch from '/MarkLogic/jsearch.mjs';
var res = jsearch
  .documents()
  .where(cts.collectionQuery('Transactions'))
  .slice(0, 10000)
  .map({extract: {paths: columns.map(x => '/' + x)}})
  .result('iterator');

  res

The result is return 4-5 times faster.

I want to use optic because I do some joins before select.

You do not mention how many items you have in your "columns" variable. 1? 100?

Whatever number that is, then multiply it by 100 *your limit). xPath has to go into the data and often extract it from compacted content.

For what you show above, I would suggest just pre-extracting your content into a view using a TDE template.

Your sample does not make use of any of the items from op.fromSearchDocs such as score for ordering etc, so in that case, you could just use:

op.fromView(your view info here)
.where(cts.collectionQuery('xxx'))
.limit(100)
....