What is the best database alternative?

[cl_integer]

I am collecting data (alarm & counter) from several serial devices which have memory and RTC.  Therefore I have to take into account that the collected data will not be chronological and I may need to sum newly collected data to existing records and then overwrite the record.  It is possible to reach about 50,000 records/month.  In addition, I need to add the data together and graph hourly, daily, weekly, monthly, etc.  So read speed is very important for user interaction.

Can someone recommend the best file type or data management practice for this?  I am using LV 8.5, so would like to use TDMS but cannot overwrite records in this version.  Currently I've settled on MGI Read/Write Anything, but is this the best solution?  I'd like to avoid DB Connectivity Toolkit and don't think I'd get any speed advantage anyway.  I just need the best way to pull out an individual record, overwrite it, and scan quickly through the entire file.

Are there any other tools in later LV versions other than overwriting TDMS records?

Clint

[Ranjeet_Singh]

Go for SQL server

[MarcoMauri]

You could consider using MySql + Native LabVIEW TCP/IP Connector for mySQL Database (https://decibel.ni.com/content/docs/DOC-10453).

Regards,

Marco

[kausreloaded]

Go to TDMS(Technical Dtata Management System) . This is a typical LabVIEW file format and works best when you have huge amount of data to handle.You can even import this data to excel and do calculations as required by you.You can go for NI DIADEM Software if you want to do more Statistical and Mathematical Data Analysis

[Schwede]

SQL

[DFGray]

It's old, but the HDF5 API will do this for you in LabVIEW 8.5.

[drjdpowell · Accepted solution]

Consider SQLite, a very fast "database in a file".  The LabVIEW wrapper I use is here; no DB Connectivity Toolkit needed.   All the stuff you describe is easy in SQL.  Write speed is greater than 100k timestamped readings per second, and read speed is higher.  You'd need to upgrade your 8.5 LabVIEW, though.

[cl_integer]

Thanks everyone.  Looks like SQL is the way to go.  Going to have a look at Damien's HDF5 recommendation, have never used it.  But I like the SQLite option.  Guess it's time to upgrade!