Batch migrating old website articles, is it more stable to import using a compressed package or a table?
The old station articles need to be moved in bulk. Is it more stable to import using a ZIP compressed package or an Excel spreadsheet? The answer depends on two things: whether the migration includes attachments and whether the content has a stable field structure. It is more stable to move the entire site with images and static files using a compressed package; for text that is mainly structured fields and needs to be mapped in bulk into articles, using a spreadsheet is more convenient. Both entry methods are supported by AnQiCMS bulk import capabilities.
Decide whether to include attachments before the migration.
Make a preliminary judgment: whether to move the images, attachments, and download files of the old station articles together. If you do, you need to retain the filename and relative path relationship, which is the strong point of ZIP compression packages; if not, only move the text and metadata, tables are more intuitive, and it is also convenient for manual proofreading. This step determines all subsequent actions, do not mix the two, otherwise, the same batch of articles may appear in two types of database paths, and it is difficult to locate the error when it goes wrong.
The applicable boundaries of the two entry methods
| Migration scenario | Recommended entry | Reason | Points to note | | Mixed text and images with local attachments | ZIP archive | Files and text stored together | Do not have too deep a directory hierarchy | | Field-standardized batch creation | Excel table | Columns directly map to content model fields | Date and null value writing should be unified | | Repeated incremental import | import/archive interface | Can query whether it exists by title | Check before writing to avoid repetition | | Old data with large structural differences | Table first and then supplement attachments | Two steps for easy verification | The second step aligns by title |
Map the column mapping before importing the table
The stability of table import mainly comes from the content model: first confirm the field list of the target model, and then decide the table columns. The common four columns are title, text, category, and tags. If there are custom fields (such as model, region, duration), the column names must correspond to the fields one by one, and temporary tags should not be inserted into the text. When exporting old data, it is convenient to do three cleaning tasks: remove the absolute addresses pointing to the old domain name in the rich text, unify the time format, and delete empty text records. If these three tasks are not done, a large amount of dirty data that needs to be modified item by item will appear after import.
How to arrange the order of compressed package import
Suggested to organize the import of compressed packages in batches: one batch corresponds to the content of a column or a quarter, and the file names within the package retain the original identifiers of the old site (such as article id or slug) for easy verification after import. Do not make the package too large, and only roll back a small batch if individual files fail. After the import is completed, first take a sample to check three types of pages: text with images, download pages with attachments, and model pages with custom fields, as these are the most prone to problems during format conversion.
Perform verification with the interface before re-importing.
The most likely to occur during cross-batch relocation is re-injection. AnQiCMS provides an interface for importing document content /api/import/archive and supports querying whether the document already exists by title, making 'first check then write' a fixed process: before each import, query the title once, and if a match is found, manually judge whether to overwrite or skip. For old sites that need long-term synchronization, this entry is more stable than repeatedly manually exporting tables.
Contingency and inspection items after migration failure
The formal document deletion will not physically delete the content, it will be moved to the recycle bin and can be restored, so when the batch is mistakenly exported, remove it first and then check, don't worry about permanent loss. At the end of the process, check four items: whether the total number of articles is consistent with the old site list; whether the images on the sampled pages can be displayed; whether the categories and tags fall into the target columns; whether the redirection mapping from the old address to the new address has been configured. Confirm that everything is correct before handling the inclusion-related actions, to avoid making changes to two layers at the same time during the migration period.
Common questions
What to do if the old site export format is not standardized?
Perform a one-time cleaning script, unify the format into two products: table columns and attachment directories, and then separately proceed with table import and compressed package import. Do not expect the import end to tolerate errors.
How many articles can be moved at one time?
Batch by category is more stable; control each batch to a size that can be manually sampled, so that when problems arise, they can be located to a batch rather than the entire site.
Will the ranking drop after the move?
The main risks during the migration period are changes in addresses and missing pages. Maintain a stable new address structure, redirect old addresses one by one, and then conduct sampling verification against high-traffic pages of the old site. Fluctuations are usually controllable.