MongoDB to DynamoDB

Attempted to convert rather large mongo dump to dynamo.

Seems like AWS data pipeline is most elegant way to do this if one can workaround with dynamo non-standard json format issue.

Get and create a collection

export json and run some sed to trim

mongoexport --db=wiki --collection=wiki -f title,text > output.json
sed -i 's/{ \"$oid\" : \"/"/g;s/\" }, \"/", "/g;s/{{{//g;s/}}}//g;s/{{//g;s/}}//g;s/$/,/g;' output.json

split file to avoid OOM, adjust 30000 based on your free ram then add [ and replace trailing, with ]

split -l 30000 output.json
sed -i '1s/^/[/;$s/,$/]/' x*

import & make a batch process to do all sliced files [bash] dyngodb2 <‘xaa’)) [/bash]

dyngodb2 <'xaa'))
  "operation": "echo",
  "TableName": "Thread",
  "Key": {
      "title-index": "!"