I don't understand the question either. A diffraction pattern does not have purity as one of its describers. Diffraction patterns cannot have unwed sex or eat cows or any of a number of things that could be considered lacking purity.

Probably, the question is not well done: if you mean to know the phases contained in your sample and detected in the XRD pattern you should perform a qualitative analysis by using a diffraction database, then you can perform a Rietveld refinement if the structure of the phases you found are known. Or maybe you need to perform an ab initio structure determination of a new phase, but this is a more complex procedure....

I agree with the poor clarity of your question. Anyway, as already noted, you should identify the crystalline phases by using a database of X-ray powder diffraction data, e.g. visit the site http://www.icdd.com/ for info . BUT, remember that your sample could contain also amorphous (non crystalline) phases; in this case you need other information, like chemical analysis, some type of spectroscopy …

it is true that the question is not well frased but it is likely that you ment how can you determine if the sample that you analized by PXRD is pure or not. 

So take a calculated PXRD pattern of your product and compare it with the PXRD pattern that you have measured. If the two match you can say (qualitatively) that you product is pure. If you observe other peaks you have impurities. A quantitative analysis must be done using the 

I think that Sarkar want to mention the intensity of peaks in XRD pattern are too low and look like noise or "not purity"? In this case, amount of sample did not enough for XRD check or that is characteristic of sample or the sample exists many phases. First, you increases amount of sample and search the data base (JCPDF cards).