numeric, date, time, etc. Formats, for its part, determine how values are displayed; for
instance, how to record dates or names of persons.
To reduce errors, it is common to ban letters in phone numbers, define the length of the ID field,
set maximum or minimum values, etc. If the database is designed with these constraints, it is not
possible to enter data if it does not comply with rules. This is important to consider when
analyzing the target database.
Understanding these dimensions is relevant to plan the migration because the process
could magnify data quality flaws. The inadvertent presence of duplicate records could lead
to replicating this error or deleting valuable information. The syntax rules of both, the
legacy database and the new database, are also essential for the migration, as data could
be rejected in the new system if it does not meet format, typology or range rules.
Additionally, assessing data quality is the only way to correctly estimate costs and time, and
allocate resources for the process. However, foremost is the fact that migration is
supposed to maintain data quality, so it is essential to identify quality baselines and goals.
Frequently, data quality is also related to external features, such as accessibility,
trustworthiness and security, which also constitute common reasons for database migration.
●
Accessibility refers to the availability of different pieces of data for those who need
it when they need it. As mentioned, this is a recurring goal when migrating offline
databases to online platforms.
●
Trustworthiness depends on the whole data management flow, but regarding data
migration is achieved by carefully documenting every transformation in the dataset
and who is accountable for each action.
●
Security is a major reason for database migration in the human rights field. It is
common to keep high security standards in the source database and plan to keep or
increase them in the target database. However, security is sometimes forgotten in
the intermediate steps, when we generate backups. These intermediate files, as well
as the retiring systems, should be wiped out once the content is verified at the
target database. As always, the best way to reduce security threats is addressing
these challenges at the planning stage.
6