It appears to be there's limitation in pydatastream DS.fetch() command. In DS.fetch([tickers],[fields]) command, (#of tickers)x(# of fields) are limited to 100, that is 10 tickers for 10 field, or 20 tickers for 5 fields, etc. I checked with pydatastream author, and confirmed that there's no restriction set in pydatastream. I am wondering if this restriction is set by Datastream server or any other reason for that limitation. Number of time-series data point does not appear to be limited.