Migrate from SQL Server to MatrixOne¶
Assess schemas, keys, types, Unicode, time, T-SQL, and privileges. SQL Server was not connected here; these mappings need validation and do not promise lossless automatic conversion.
Types and objects¶
SQL Server |
Candidate/policy |
Checks |
|---|---|---|
bit |
BOOLEAN or tested 0/1/NULL representation |
MO has boolean capability; verify NULL |
real / float |
FLOAT/DOUBLE by source precision |
Approximate/special values; not exact currency |
money / smallmoney |
DECIMAL retaining four fractional digits/range |
Source extrema and arithmetic rounding |
datetime / datetime2 / datetimeoffset |
DATETIME plus original value/precision/offset columns |
Seven digits become six-digit precision in DATETIME(6); define rounding/truncation; retain offset rather than merely renaming to timestamp |
nchar / nvarchar / ntext |
CHAR/VARCHAR/TEXT |
Unicode, length units, collation/trailing spaces; no global N-prefix removal |
binary / varbinary / image |
VARBINARY/BLOB |
Original/zero bytes and encoding |
uniqueidentifier |
UUID or explicit text/byte policy |
Byte order, ordering, driver representation |
rowversion |
Retain binary marker and business policy |
Version counter, not a timestamp |
geometry / geography |
WKT/WKB + SRID and function verification |
Planar/geodesic, axes, units |
hierarchyid / sql_variant / XML |
Dedicated representations/alternatives |
Methods/query/type semantics are not automatic |
Handle identities/sequences, computed columns, CHECK/unique constraints, T-SQL programs/triggers, SQL Agent jobs, schemas, and privileges individually. Removing constraints or executing CREATE does not establish semantic equivalence.
Workflow¶
Create reviewed DDL on isolated targets and test boundaries/actual schema.
Pin exporters, DBeaver, or DataX; define consistent reads/write suspension and failed batches.
Match CSV dialect, NULL/empty values, Unicode, dates, and columns. Size batches by bytes/transactions/memory; 5000 rows does not guarantee direct S3.
Review SQL data and Unicode literal conversion; do not run
sed s/,N/,/g. Execute the actual data file instead of repeating the DDL file, and retain failures.Compare key sets, exact values, timestamps/offsets, bytes, and queries. Live migration needs validated SQL Server CDC/log positions, retention, consumption, and cutover.
See migration/backout workflow and target numeric/spatial experiments. SQL Server, exporters/CDC, and DBeaver/DataX pipelines were not executed.