storage problems in RDBMS
The growing amount of unstructured data presents a problem for relational databases.
The rows and columns of a relational database are ideal for storing sets of values,
but most information is composed of much more than that. Consider something like
a person’s medical record. It is incredibly heterogeneous. It includes values
(name, date of birth), relationships (to family members or care providers,
to symptoms and medications), geospatial data (addresses), metadata
(provenance, security attributes), images (CAT scan), and free text
(doctors’ notes, transcripts).
Now, imagine putting all of that data into a Microsoft Excel spreadsheet…
That is exactly what it’s like trying to put this type of data into a
relational database. It’s not easy and it requires a lot of ingenuity,
and many difficult choices.
RDMS was not designed for heterogeneous data.