If the table has data for a past year, and you substitute all of it with data for a later year, then this is classified as new data. No acknowledgement is required, although it is polite to the reader to include a reference to the previous data, in case the reader wishes to study trends.
However, if you omit columns, rows or cells as part of your adaptation, then a full reference is demanded. Even correcting an incorrect entry or modifying a heading requires you to fully cite the original table, because you did not collect the original data.
If your review augments the table and adds columns or rows from another source, this too must be cited.
If the table is unique in any other way, such as being a new instrument, then you must of course reference the original.
Since the manuscript is a review article, copying data tables from some published sources for comparison purposes is OK. In your case, you said that you updated the data presented in a previously published table. In this specific case, no plagiarism is performed provided that previous references are correctly cited. Actually, showing the old data (or results) against the updates is necessary for the reader to understand the changes.
An acknowledge is required, whether you use part of all of the article, and the acknowledgment should indicate what changes (omissions, additions, etc) have been made to the original, and in what ways, if any, the original has been altered.