eriky / esclient Goto Github PK

Hello,

# esdump -h
Traceback (most recent call last):
  File "/usr/local/bin/esdump", line 7, in <module>
    import esclient
  File "/usr/local/lib/python2.7/dist-packages/esclient.py", line 1, in <module>
    import requests
ImportError: No module named requests

After this, I installed easy_install + pip. Then removed it with "pip uninstall esclient".

At this point, installing with "pip" it worked.

I am currently running an es dump and there is no progress i know u cant let me know the % done but would be useful to at least know how many documents have been written already.

You can also grab the total number of documents at the start and then use that as a guildline and then every so often make that call again and readjuest the number of percentage done. you could also wait to the end and just show a number bigger then 100% if it was not for disk io monitoring and watch ls -la i would not know if it was working or my connection froze

Hi,

First of all - good work, this plug in really works.

The only issue i see is the Import takes time. I had an index just of 1k size document each, total 273781 document and when i imported it - the time consumed was around 20-30 minutes. The size of the whole index is around 50MB.

We are planning to do backup of 50GB or 100GB data, and based on above it will be like day's when the import completes.

Any hints or ideas appreciated to resolve this?

thanks
pranav.

Implement bulk indexing to speed up large indexing jobs

pip install esclient did not download argparse

Since it dependence on Response.text, it would be cool if the minimum version could be set to 0.10.0, where this method was introduced.

Hi, one obvious problem is the lack of support for encoding custom types... for example datetime.datetime currently fails saying its not JSON serializable.

The best way to do this would be to pass an encoder class to the json.dumps() call in send_request. I was thinking if you could supply an encoder and decoder class when you call the ESClient constructor it would give the maximum flexibility, but still keep the ESClient API simple.

Splitting the library into two parts:

• the ElasticSearch API
• a higher level library that makes it even easier to use ElasticSearch

The higher level library could, for example, contain methods to:

• do a query of type query_string without having to manually create the search options.
• change replication settings for indexes with a single api call

Traceback (most recent call last):
File "/usr/local/bin/esdump", line 66, in
scroll_id = es.scan(query_body = query_body, indexes = indexes)
File "/usr/local/lib/python2.7/dist-packages/esclient.py", line 235, in scan
return result["_scroll_id"]
KeyError: '_scroll_id'