Data migration to PostgreSQL

Every record counted, every amount compared.

From VSAM and DB2 to an open database.

Legacy data sits in VSAM files, DB2 or IMS databases and sequential files, with EBCDIC encoding, packed decimals and fields that carry three meanings. We model it in PostgreSQL, migrate in trial runs we can repeat identically each time and deliver a reconciliation report with every run.

Start the conversation →See the diagram →

What you can expect

Repeatable

Scripts, not manual work. The migration on cutover day is the same one as in the trial runs.

Reconciled

Record counts, control totals on amounts and samples per table, source against target.

Cleaned where you decide

Data quality issues show up in a report. You decide what gets corrected and what stays.

From source to table

Legacy data structures and their shape in PostgreSQL
Source	PostgreSQL	Point of attention
VSAM KSDS	Table with a primary key	Alternate indexes become indexes
REDEFINES and OCCURS	Separate tables or typed columns	Every variant gets an explicit shape
COMP-3 and zoned decimal	NUMERIC with precision and scale	Precision is preserved
Dates as PIC 9(6) or 9(8)	DATE or TIMESTAMP	Invalid and zero dates get an agreed rule
EBCDIC text	UTF-8	Code page recorded per source
IMS segments (hierarchical)	Tables with foreign keys	Parent-child relations become constraints
Generations and history	History tables or archive	Retention periods agreed up front

How we do it

01Preparation

Data profile

Per field: fill rate, value range and exceptions, measured on the real data.

Mapping document

Every source field has a target column and a conversion rule, validated by you.

02Execution

Trial migrations

Full runs on a copy of production, as often as needed until reconciliation closes.

Cutover window

Its duration is measured in the trial runs, not estimated.

03Control

Reconciliation report

Per table: counts, control totals and rejected records with the reason.

Fallback plan

The source stays untouched until the cutover is formally accepted.

■PostgreSQL · own servers or cloud

■Repeatable migration scripts

■Reconciliation report per run

■Data stays inside your environment

Curious what is inside your COBOL?

We start with an inventory.Start the conversation →