Just wondering if anyone knows what database format DataHQ or Premium Data will be using in future (say 2-5 years). Anyone got any information on this?
They currently use the Metastock Database format which has been around since I believe 1985. But now that Metastock have dumped their own database after nearly 30 years, do these data providers have any future plans for an alternative?
It's an interesting topic because sooner or later 64 bit platforms will be the norm and these data providers will have to maintain the Metastock Database using 64 bit applications to keep up with the times. Do they have 64 bit applications for this?
A Metastock database contains a master, emaster and xmaster file, which all contain the names of the stocks and an index to corresponding .dat or .mwd files among other things. The .dat and .mwd files store all the quotes for a stock in microsoft binary floating point format. On top of this, I believe only 65535 quotes can be stored, which is still plently for EOD data. For EOD data, I believe a quote record currently contains DATE, OPEN, HIGH, LOW, CLOSE, VOLUME & OPEN INTEREST.
One minor problem with storing data in floating point format is there could be a slight loss of accuracy in some fields. For example the Volume of some highly traded US stocks could lose a precision if it is over I think 10 million or 100 million due to only being able to store 5 or 6 significant digits, but practically speaking, this is not a big issue.
Metatrader history files use double precision, but there is no master file so you have to iterate through each file to get a symbol list. What genius format will these companies use and how will they handle share splits and what not if and when they decide to change database formats?