Commercial deck: View the presentation · Get the PDFs
What NSAdaDB is
nsadadb copies an Adabas file to a relational or document database, verifies that source and destination hold the same data, and syncs only what changed. Three commands:
copy— initial load (merge,empty,dropmodes).compare— read-only diff report (equal / missing / changed / extra). Exits0when equal,1on differences.sync— applies only missing and changed rows; with-delete-missingit also deletes destination-only rows.
Destinations: sqlite · postgres · mariadb (mysql) · mssql (sqlserver) · oracle · mongo (mongodb) · elasticsearch · cassandra (scylla) · kafka (copy only; db2 unsupported).
Database numbers, files, fields and paths in the examples are fictional.
Use cases
1. Operational replica on SQLite
Initial load with table recreate, then nightly diff sync.
nsadadb copy -dbid 50 -fnr 15 -kind sqlite -path ./replica.db -table CUSTOMERS -mode drop
nsadadb sync -dbid 50 -fnr 15 -kind sqlite -path ./replica.db -table CUSTOMERS -delete-missing
2. Migration to PostgreSQL with DDM names
Copy through the logical dictionary with a target file so the DSN never hits the shell history.
# dst.yaml
kind: postgres
dsn: postgres://user:pass@host:5432/demo?sslmode=disable
table: CUSTOMERS
nsadadb copy -dbid 50 -config ddm50.json -ddmname MY-FILE -target dst.yaml -mode drop
nsadadb compare -dbid 50 -config ddm50.json -ddmname MY-FILE -target dst.yaml -json
3. Publishing to Kafka for the data lake
Every ISN travels as JSON keyed by ISN (deletes are tombstones).
nsadadb copy -dbid 50 -fnr 15 -kind kafka -dsn kafka://host:9092/adabas -table CUSTOMERS
4. Auditable reconciliation
compare with a JSON summary proves source and destination match, ISN by ISN.
nsadadb compare -dbid 50 -fnr 15 -kind oracle -dsn "oracle://u:p@host:1521/svc" -table CUSTOMERS -json -outfile diff.json
5. MU/PE in child tables
With -mu-pe child, groups normalize into <table>__<group> with PK(isn, idx); scalar MUs map to a VALUE column. Deletes cascade.
nsadadb copy -dbid 50 -fnr 15 -kind postgres -dsn "..." -table CUSTOMERS -mu-pe child -mode drop
Option reference
| Option | Description |
|---|---|
copy|compare|sync | Command to run (first argument). |
-dbid | Source Adabas database. Required. |
-fnr / -ddmname -config | File by number or by logical DDM name. |
-kind, -dsn/-path, -table | Destination, connection and table/collection/index/topic. Or all in -target YAML. |
-fields, -search, -sortby | Source selection (no -fields with -mu-pe child). |
-fromisn, -toisn, -limit | ISN window and record cap. |
-mode merge|empty|drop | copy mode (default merge). |
-delete-missing | Also handle destination-only rows (skipped with a filtered source). |
-mu-pe json|child | MU/PE groups as JSON (default) or child tables (relational only). |
-batch, -workers | Rows per bulk write (500) and parallel workers (4). |
-state, -resume | Checkpoint every 5000 records and resume. |
-dry-run | Preview without writing. |
-json, -outfile | Machine summary and detail file (up to 10,000 lines). |
-multifetch, -blocksize | Adabas read tuning. |
-silence, -debug | Silence and diagnostics. |
Exit codes
0— success (compare: same data).1— parameter error, orcomparefound differences.2— configuration, connection or apply error.
License
Per-product signed license (app=nsadadb) at $NS_HOME/licenses/nsadadb.key
(mode 0600). Without a valid license the binary does not run; there is no
automatic trial (a trial is a signed license issued with nslm issue --trial).
-license-status shows the status, validity and this application’s fingerprint
on this machine. Demo builds stop after 50 records.