@Tiantian_Wang: More details for what @gbbio is explaining can be found in the FAQs here: https://galaxyproject.org/support/
- Format help for Tabular/BED/Interval Datasets
- Common datatypes explained
- The tool I’m using does not recognize any input datasets. Why?
- How do I find, adjust, and/or correct metadata?
- My job ended with an error. What can I do?
Galaxy makes an attempt to autodetect datatypes but that can sometimes be a challenge. If a file has a format or content problem that makes it “out of specification” for an expected datatype assignment, you might need to make some adjustments (remove extra headers, etc) for the dataset to be successfully used with tools to produce usable analysis results.
The metadata “datatype” assignment is an interpreted attribute – and some tools are more forgiving than others when the data itself deviates from strict formatting expectations. And it is certainly possible to directly label data with a mismatched datatype.
Analysis jobs with actual problematic inputs would fail/error or produce odd results, even when running the underlying tool command-line (not just in Galaxy). The FAQs above plus information on tool forms themselves can help to resolve issues like those.
I added a few tags to your post to show the range of format issues discussed in prior Q&A. This is a good one – although it seems unlikely that you would have the exact same format problem – it may help anyway: Galaxy cannot recognize csv files
Thanks!