Allow case sensitivity in metadata columns

[josenavas]

e.g. alleles

[antgonza]

This is allowed in pgsql http://stackoverflow.com/questions/20878932/are-postgresql-column-names-case-sensitive, the question is: do we really need it? I think in a analysis point of view, we do not, as having a column with different values is much better than having multiple with just yes/no.

[rob-knight]

Yes we need it. Typical use case as previously noted is alleles in mice: AA, Aa, aa for the different genotypes resulting from an Aa x Aa cross for example, critical for generating matched litter mates and critical to distinguish.

On Oct 16, 2014, at 1:16 PM, Antonio Gonzalez <notifications@github.commailto:notifications@github.com> wrote:

This is allowed in pgsql http://stackoverflow.com/questions/20878932/are-postgresql-column-names-case-sensitive, the question is: do we really need it? I think in a analysis point of view, we do not, as having a column with different values is much better than having multiple with just yes/no.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/439#issuecomment-59414789.

[ackermag]

Also needed for host_scientific_name

[antgonza]

I believe there is a confusion between column names and values. Currently,
column names are not case sensitive but values are, which is the common
postgres behavior. My point is that we do not want to have columns Aa and
aa with column values TRUE/FALSE but we want to have a (or several) column
called alleles with values Aa/aa. IMOO the former case makes analysis more
difficult and we always tend to combine information so
it reassembles the latter.

[rob-knight]

Well, then you get into flow cytometry where for every allele for every sample you might have counts of how many cells you saw in a given antigen-labeled fraction. Etc. I wouldn't bet there aren't other use cases so I recommend against making these analyses impossible.

On Oct 20, 2014, at 2:41 AM, Antonio Gonzalez <notifications@github.commailto:notifications@github.com> wrote:

I believe there is a confusion between column names and values. Currently,
column names are not case sensitive but values are, which is the common
postgres behavior. My point is that we do not want to have columns Aa and
aa with column values TRUE/FALSE but we want to have a (or several) column
called alleles with values Aa/aa. IMOO the former case makes analysis more
difficult and we always tend to combine information so
it reassembles the latter.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/439#issuecomment-59706082.

[antgonza]

Makes sense. Thanks.

[ElDeveloper]

👍