I found that the structure for TiFeSi is given differently in ICSD and Pearson crystal database as follows:
The Wykoff positions are given in the Pearson database (data set no 1822291)
Ti1 Ti 4 b 0.25 0.2207 0.0206
Ti2 Ti 4 b 0.25 0.4979 0.1677
Ti3 Ti 4 b 0.25 0.7996 0.0463
Fe1 Fe 8 c 0.5295 0.1236 0.3699
Fe2 Fe 4 a 0 0 0.0
Si1 Si 8 c 0.506 0.3325 0.2452
Si2 Si 4 b 0.25 0.0253 0.2554
The Wykoff positions are given in the ICSD database (database code 41157)
Ti1 Ti0+ 4 b 0.25 0.2004(7) 0.2964(14)
Ti2 Ti0+ 4 b 0.25 0.7793(6) 0.2707(14)
Ti3 Ti0+ 4 b 0.25 0.9979(6) 0.9178(15)
Fe1 Fe0+ 8 c 0.0295(7) 0.3764(4) 0.12
Fe2 Fe0+ 4 a 0 0 0.2501(12)
Si1 Si0+ 8 c 0.0060(13) 0.1675(9) 0.9953(18)
Si2 Si0+ 4 b 0.25 0.9747(11) 0.5055(23)
Although the lattice parameter for both of the database is almost the same.
Which one should I take for ab initio calculations or XRD Rietveld refinement?