Normalizing CSV to MySQL? - mysql

Question

I'm new to the whole "normalized table" thing. I have a csv file with the contents as follows:
Cell,Width(m),Length(m),Spacing(m),VDD(V),VSS(V),Temp,Param,Value,Path,TOOL
pmos_var12,5e-03,5e-03,5e-03,0.5,0,0,delay[s],4.65e-06,/home/user/tests/run2/pspice
pmos_var12,5e-03,5e-03,5e-03,0.5,0,10,delay[s],6.2e-06,/home/user/tests/run2/pspice
pmos_var12,5e-03,5e-03,5e-03,0.5,0,25,delay[s],7.46e-06,/home/user/tests/run2/pspice
pmos_var12,5e-03,5e-03,5e-03,0.5,0,70,delay[s],8.98e-06,/home/user/tests/run2/pspice
pmos_var12,5e-03,5e-03,5e-03,0.5,0,100,delay[s],9.56e-06,/home/user/tests/run2/pspice
nmos_var12,5e-03,5e-03,5e-03,0.5,0,0,delay[s],4.65e-06,/home/user/tests/run2/pspice
nmos_var12,5e-03,5e-03,5e-03,0.5,0,10,delay[s],6.2e-06,/home/user/tests/run2/pspice
nmos_var12,5e-03,5e-03,5e-03,0.5,0,25,delay[s],7.46e-06,/home/user/tests/run2/pspice
nmos_var12,5e-03,5e-03,5e-03,0.5,0,70,delay[s],8.98e-06,/home/user/tests/run2/pspice
nmos_var12,5e-03,5e-03,5e-03,0.5,0,100,delay[s],9.56e-06,/home/user/tests/run2/pspice
I've created these tables to store the data:
CREATE TABLE `TEST__RUN_MAPPING` (
`ID` int(11) NOT NULL auto_increment,
`NAME` varchar(45) NOT NULL,
`STATUS` varchar(20) NOT NULL,
`PATH` text NOT NULL,
`TOOL` varchar(10) NOT NULL,
`COMMENTS` text NOT NULL,
PRIMARY KEY (`ID`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1
CREATE TABLE `TEST__DATA_MAPPING` (
`ID` int(11) NOT NULL auto_increment,
`NAME_ID` int(11) NOT NULL,
`CONDITIONS` int(11) NOT NULL,
`VALUE` varchar(10) NOT NULL,
PRIMARY KEY (`ID`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1
CREATE TABLE `TEST__CONDITION_MAPPING` (
`ID` int(11) NOT NULL auto_increment,
`CELL_ID` int(11) NOT NULL,
`W_ID` int(11) NOT NULL,
`L_ID` int(11) NOT NULL,
`SPACE_ID` int(11) NOT NULL,
`VDD_ID` int(11) NOT NULL,
`VSS_ID` int(11) NOT NULL,
`TEMP_ID` int(11) NOT NULL,
`PARAM_ID` int(11) NOT NULL,
PRIMARY KEY (`ID`)
) ENGINE=MyISAM DEFAULT CHARSET=latin1
TEST__RUN_MAPPING ID maps to TEST__DATA_MAPPING NAME_ID
TEST__DATA_MAPPING CONDITIONS maps to TEST__CONDITION_MAPPING ID
All *ID in TEST_CONDITION_MAPPING map to their own table in order to have things unique.
each one of these csv files will differ in what technology used in the simulations, and I keep tabs on this with the NAME column in TEST__RUN_MAPPING. Cell, Width(m), Length(m), Spacing(m), VDD(V), VSS(V), and Temp are all values that are swept, but usually they're the same per technology so I grouped them together in a separate table.
Are there any other ways that a more experienced person could break down the relationship such that it can have optimal reading times? better normalization?

Answer 1

If I understand this structure correctly then I would not seperate the conditions table from the run table. Surely they would have a 1-to-1 relationship. So why not have them both in the same table. The conditions for that paticular run.
Also I would be careful about putting a 'TEXT' block inside a record. TEXT and Blobs can cause some performance problems. varchar in mysql 5 can go as high as 65,000 characters. For paths you should not need more than 1024 characters. So varchar(1024) should be enough for a path.