Re-import from the broker, not from the journal
The instinct is to export from the old tool and import into the new one. Better: go back to your broker and export the raw history again.
Broker exports are the original record — full precision, every trade, the costs as itemised, and on MetaTrader the stop and target. A journal's own export is a processed copy of that, often with columns dropped and figures already combined. Starting from the source avoids inheriting whatever the previous tool decided to round, merge or omit.
How to export from each platform.
Import both if you are unsure, into separate accounts, and compare the trade counts before deleting anything.
What moves cleanly
Instruments, direction, size, times, prices, results, and costs where the broker itemises them. This is most of a journal by volume and essentially none of it by value — it is the half a machine can always reconstruct.
What usually does not
Notes. Sometimes exportable as a column, often not, and rarely in a form another tool will attach to the right trade.
Screenshots. Almost never. They live in the old tool's storage, and the export contains links to it.
Tags. Technically portable and practically the hard part — see below.
Anything computed. Ratings, scores, streak counts. These are outputs of the old tool's arithmetic, and importing them into a system that computes differently produces two incompatible numbers with the same name. Let the new tool recompute from the trades.
Tags are the real work
A tag vocabulary is a private language. late-entry means something specific to you, and by trade three hundred it has quietly come to mean two different things — the trade you took after the level was gone, and the trade you took after hesitating. Only you know which entries are which, and the export carries only the word.
Which makes a migration the one good moment to do something you would otherwise never do: shrink the list.
Look at how often each tag was used. The ones applied twice are noise; the ones applied to a third of your trades are too broad to group anything. What survives should be a short list where each name would make you enter or skip a trade differently — the same test as for journal fields.
Map the old vocabulary onto the short list as you import, and accept that some history becomes untagged. Untagged is honest; mis-tagged is worse than nothing.
Check these three things after importing
Trade count against the old tool. A large difference on stock or crypto history usually means one of them counted fills rather than round trips — see ledgers.
Total P&L against your account statement, not against the old journal. If they differ, the usual cause is costs counted twice or not at all — see net P&L.
Open positions. Trades opened before your export window have no entry to match against and can vanish or appear stranded.
Keep the old one until you have checked
Two weeks of overlap costs one subscription and settles every discrepancy while both sets of numbers still exist. Cancelling first is how a mismatch becomes permanent — you get one number, no way to reproduce the other, and no idea which was right.