"paingirahasyabrahmama" is a good example. Most of it is claimed to be lost - but AFAIK, only a few lines supporting one sect are supposed to have survived.
Is this a fair way to search for scriptural knowledge?
For example, partial readings of what we know of the Mahanarayana Upanishad can lead to widely different conclusions.
Narayana is particularly solemnized in the 11th chapter of the text, calling Atman (soul) as Narayana. This description mirrors those found in Yogashikha Upanishad and Yogatattva Upanishad. Narayana is described as the highest goal, the light beyond, the highest self, the highest Brahman, the highest object of thought.
The chapter 12 and twenty six verses that follow then solemnize Rudra, in a manner similar to Narayana, as being all the universe, the manifest One, the right, the just, the truth and the highest Brahman. Once again, the text references and integrates numerous hymns and their fragments from the Vedas, as it solemnizes Narayana and Rudra.
there is the risk of the "blind men and an elephant" fallacy.
For ancient texts, there might be lost portions, interpolations, natural distortions in transmission etc. - but we need a significant amount of material to be able to draw conclusions from what is still available.
An enormous corpus is still preserved in the case of Shruti works - not one or two lines making sharp claims as in the case of paingirayasyabrahmana.
EDIT: clarified to focus the question.